Home

CanariCam User Manual

1. unis 30 ZU HON ECCE SONT NPOROEREREES 3 AI Stc ouem cha sink Dcum Del eat UE 32 GTS ETT DR ae ese 32 4 4 CanariCam Spectroscopy Data Reduction sese 32 Reduction of 36 24 2 INWMIGL PLOSOSSID usse ee tetucddtitssenadeun ____ 37 EE 37 4 4 4 Wavelength Transformation 4 44 9 SDOGLFUL EXIFOCHON a cds Rotten i vo o ide quitte a ie 42 O OPE tan dotar t tate 43 44 7 Spectral Flux COllDEOlOn usta qe n tata cta nite ad nee 46 4 5 CanariCam Polarimetry Data Reduction eese 49 Zo Pata kedun tion doi obs 52 _ iv Version 1 2 2009 04 23 2 DIDIT NIU Rm 53 4 5 CanariCam Coronography Data Reduction ccccccccccccccceeeeeeeeeeeeaeeseeeeeseeeees 55 Appendix A s ees tet ein dtes Un ast en aed lust educ dtes Ua ao qd 56 E DIEE 56 II Backeroutid tatu eem rM Manius 57 ILE The Standard Chop Nod Technique FREIE vx teet Nod eee ron 60 H I Comparison of Signal to Noise for 5 61 Version 1 3 2009 04 23 1 Instrument Introduction
2. vitii ea Pista VALUE SR RU a uite aude 10 2 J nstt tietit OP S d deb betta ouo a edd bertus 11 2 4 Dewar Entrance Wi GOW A 13 2 PECO W NEE MERERI 16 2 9 MaW ave Plate Wee Macon T 17 ADEREN 18 2 oot n ded tudo ethene 19 ZO S NW SAG A oa ete eee 19 2o MM Vy DIOS si ast ete ck paar la E bal dd acta tet DE 20 2 8 Wollaston Prism She vod tata ad 21 2 9 Mirror amp Grating TUELEeL eoe epe t ees e e aste x deu eae vta oe a edt 21 3 Software Operation Queue and Classical 24 4 Observing and Data REGUCH Oils uude vete nee prete eu UE ede re dX IHE eae 24 Zu MIAO CUO TION occi a ote E nS ive alus etus wc ive ada ten E 24 42 NOG CNG ac Cote eut re e totom al dodo uestes 25 4 3 CanariCam Data Structure amp Exposure 25 4 3 CanariCam Filters and Calibration Objects sse 26 4 3 CanariCam Imaging Data nennen 29 40d DLBACRGROUNMNDS iens Ea E eie utet m nad ep nie ini a eiua is 29 40 450 205 ne ei 30 o
3. anari Arm Astronomers User Manual Version 1 3 23 April 2009 Prepared by Chris Packham and the CanariCam instrument team Charles Telesco David Ciardi Robert James Radomski Francisco Reyes Dan Cawlfield Jim French Kevin Hanna Roger Julian Jeff Julian Margaret Moerchen Frank Varosi Version 1 3 2009 04 23 Preface The purpose of this manual is to provide information regarding the use of CanariCam the mid IR multi mode camera for the Gran Telescopio Canarias GTC CanariCam is a mid infrared 8 25 um imager long slit spectrograph dual beam polarimeter and coronagraph built by the University of Florida Infrared Astrophysics Group for use at the GTC by the astronomical community In this version of the document certain assumptions are made regarding the system capabilities that are subject to change during commissioning Specific questions about the system should be directed to the GTC Help Desk The CanariCam instrument team is heavily indebted to the T ReCS instrument development team as listed below as well as the Gemini observatory especially those persons listed below T ReCS Instrument Development Team Charles Telesco Robert Pifia Chris Packham David Ciardi Francisco Reyes Naibi Marinas James Radomski Cynthia Gomez Abdulatif Albusairi Dan Cawlfield Jim French Kevin Hanna David Hon Roger Julian Jeff Julian Tom Kisko Frank Varosi Gemini Tom Hayward Kevin Volk Rachel Mason Scott Fi
4. At low spectral resolution B A F and G type stars have a smooth spectrum in the mid IR region and K and M type stars should be avoided For a list of B G type stars see the compilation at the Gemini mid IR resources WWW site http www gemini edu sciops instruments mir MIRSpecStdBAFG html and a screen shot of which is shown below 33 Version 1 2 2009 04 23 omini Bright Stars with Spectral Types B G A list of bright telluric standard stars of spectral types B G for calibration of Michelle spectra is provided below along with estimated in most cases magnitudes at N and Q Note the following restrictions on use of the stars in the list below At low resolution B A F and G stars have smooth spectra in the 7 25um region and therefore all may be used to correct for atmospheric absorption features At moderate resolution G stars also should be avoided but B F stars are OK At high resolution only B and early A stars are suitable except for measuring hydrogen lines All of the stars in the list are bright enough to obtain good S N lowN grating spectra in reasonably short integrations For 1 spectra only the stars brighter than 151 magnitude are bright enough For medN2 and lowQ spectra only the stars brighter than Oth magnitude are bright enough For echelle spectra only Sirius BS 2491 is bright enough asteroids almost always must be used HR BS RA 2000 Dec V Sp N amp Q est zz existing phot
5. ISAZA VAS Nearest 3 HD79554 pos 09 15 13 85 14 56 29 4 V 5 360 12 5 492Jy Nearest 9 HD81146 pos 09 24 39 26 26 10 56 4 wV 4 473 12 9 826Jy Nearest 4 HD85503 pos 09 52 45 82 26 00 25 0 V 3 88 12 14 630Jy Nedrest 205 BDH979B 1022419 TA AL Z9toH ep VEA 066 EZISLON HRA Nearest 5 HD74442 pos 08 44 41 10 18 09 15 51 3 94 1121 11 2407 gt gt gt o wa 221 2 Done Abok 21 lt gt Figure 12 Gemini tools for flux standard selection PSF standards are observed to monitor the image quality of the telescope We note that sometimes the flux standard can be used for PSF calibration but that is the exception rather than the rule since as note above in many cases the flux standard is too bright too blue or an unrepresentative color compared to the science source The best PSF sources are giant stars of type K or M Note that M supergiants are usually avoided due to possible extended dust shell emission although they are sometimes useful secondary standards Also very bright stars should be avoided due to the array artifacts they cause PSF stars can also be selected by using the Hipparcos on line database When using PSF standards and depending on the program goals one may need to take into account pupil rotation inherent in any long time observations or when slewing to other sources the PSF widths at all brightness levels for a hexagonal
6. 5 607 12 7 451 guidestar U1200 07711013 11 6 mag Enter a target position below format as in the OT say 16 44 29 470 and 23 47 58 20 for RA and Dec Nearest 7 HD152326 pos 16 51 45 26 24 39 23 2 d 15 3490 degrees COHEN esnectivelv Alea enter the le tho a tack will retur ll dadai 5 036 12 5 238 guidestar GSC0206201039 12 21 mag respectively Also enter the length of the observation The task will return the names of the 10 closest Nearest 8 HD138481 pos 15 30 55 76 40 49 59 0 d 17 2690 degrees COHEN stellar standards one might use for photometric or low resolution spectroscopic calibration for three v 5 050 12 14 880Jy guidestar GSC0305500846 10 93 mag positions 1 the target position 2 same declination right ascension less by half the entered Nearest 9 HD127093 pos 14 28 46 03 25 51 14 0 d 17 2885 degrees COHEN exposure time and 3 same declination right ascension plus half the entered exposure time 6 726 12 10 350 guidestar U1125 06980478 11 5 mag Nearest 10 HD140573 pos 15 44 16 07 06 25 32 3 d 17 3739 degrees COHEN For example for a 2 hour exposure the task will look for standards at RA hour less and hour more D EPA NAM than the target position as well as using the target position Three sets of potential standards After standards returned Nearest 1 HD158899 pos 17 30 44 31 26 06 38 3 d 3 8801 degrees COHEN
7. 4 402 12 15 910 guidestar GSC0207900371 10 47 mag Nearest 2 HD168323 pos 18 18 07 73 23 17 48 9 d 7 7243 degrees COHEN 6 487 12 7 186 guidestar GSC0209301273 11 81 mag Right Ascention 16 44 29 470 Declination 23 47 58 20 Nearest 3 HD156652 pos 17 17 34 65 28 54 47 8 d 7 9025 degrees COHEN 6 917 12 5 634 guidestar 01125 08135093 11 3 mag BN X Nearest 4 HD169414 pos 18 23 41 89 21 46 11 1 d 9 2597 degrees COHEN length of observation decimal hours 2 3 84 12 16 940 guidestar GSC0158100449 11 35 mag Nearest 5 HD170951 pos 18 31 09 66 25 09 47 8 d 10 7022 degrees COHEN Retrieve Output Clear Form V27 89 12 724 839Jy guidestar 01125 09673932 9 2 mag Nearest 6 HD152326 pos 16 51 45 26 24 39 23 2 d 12 0488 degrees COHEN 5 036 12 5 238 guidestar GSC0206201039 12 21 mag Nearest 7 HD163770 pos 17 56 15 18 37 15 01 9 d 13 6852 degrees COHEN This page was last modified on December 28 2006 3 851 12 18 390 guidestar GSC0262000596 10 64 mag Nearest 8 HD173780 pos 18 46 04 48 26 39 43 7 d 14 2084 degrees COHEN 4 833 12 6 352 guidestar GSC0211602371 11 75 mag Fourie tee nomini Nearest 9 HD168775 pos 18 19 51 71 36 03 52 4 d 14 4457 degrees COHEN Kevin Volk kvolk amp gemini edu V 4 323 12 9 060Jy guidestar 01200 09160538 11 9 mag Nearest 10 HD156283 pos 17 15 02 83 36 48 33 0 d 14 4663 degrees COHEN 3 156 12 4
8. To extract the spectrum a cut 1s made across the differenced spectrum and then this is traced across the dispersion direction For brighter objects this 1s routine but 1n fainter objects the apall command can fail due to lack of flux One should typically set the extract of apall to yes so that the tracing can be interactively examined Usually points the to far left and right of the spectra are unreliable as typically the spectrum is spread out over only 90 of the array and hence the ends of the spectra are noise In the case of CanariCam the spectra are not curved and hence the tracing should be approximately flat 1 pixel across the detector Also the spectral tracing should be close to linear If either of these cases are not met from automatic tracing points should be excluded until a reasonable tracing 1s obtained An example of a good tracing 1s shown below and note that points were deleted to the left and right of the main spectra Tracing of the spectra can be lost at shorter or longer wavelengths due to strong and variable water bands in these regions so there 1s no absolute wavelength cut on off rather it is condition dependant In the case below as the aperture used is 6 pixels wide a linear fit would have provided just as good a fit than the curved fit shown 42 Version 1 2 2009 04 23 Figure 23 Spectral extraction curve 4 4 6 De fringing of Spectra MIR spectra can be subject to fringing especia
9. 18 39 5 d 12 0745 degrees COHEN 5 455 12 11 410 guidestar U1050_ 07684466 11 3 mag Nearest 7 HD141992 pos 15 51 15 91 20 58 40 5 12 6171 degrees COHEN 4 751 12 18 060 guidestar GSC0150200617 13 48 mag Nearest 8 HD141477 pos 15 48 44 38 18 08 29 6 d 14 1806 degrees COHEN 4 109 12 44 310 guidestar U1050_ 07645008 15 2 mag Nearest 9 HD156283 pos 17 15 02 83 36 48 33 0 14 5723 degrees COHEN 3 156 12 48 190 guidestar GSC0260401448 12 94 mag Nearest 10 HD153210 pos 16 57 40 10 09 22 30 1 14 7637 degrees COHEN V 3 20 12 24 990Jy guidestar GSC0097501622 12 44 mag Before standards Nearest 1 HD141992 pos 15 51 15 91 20 58 40 5 d 3 2268 degrees COHEN 4 751 12 18 060 guidestar GSC0150200617 13 48 mag Nearest 2 HD142574 pos 15 54 34 61 20 18 39 5 d 4 1986 degrees COHEN 5 455 12 11 410 guidestar U1050 07684466 11 3 mag Nearest 3 HD143107 pos 15 57 35 25 26 52 40 4 d 4 2695 degrees COHEN 4 143 12 13 120 guidestar GSC0203701751 11 53 mag Nearest 4 HD141477 pos 15 48 44 38 18 08 29 6 d 5 7441 degrees COHEN Finding Nearby Standard Stars for Michelle or T ReCS V 4 109 12 244 310Jy guidestar U1050 07645008 15 2 mag Nearest 5 HD149009 pos 16 31 13 43 22 11 43 6 d 10 8702 degrees COHEN Calibration 5 781 12 8 218 guidestar GSC0151801061 12 07 mag Nearest 6 HD143435 pos 15 58 57 71 36 38 37 6 d 13 2160 degrees COHEN
10. 3 24 48 79 09 01 43 9 3 62 G6III 1 5 1195 3 49 27 25 36 12 00 9 4 17 G9II III 1 9 1409 4 28 37 00 19 10 49 6 3 54 G9 5III 1 1 1464 4 35 33 04 30 33 44 4 3 82 G8III 1 5 1708 5 16 41 36 45 59 52 8 0 08 G5III N 1 94 0 1 93 1784 5 23 56 83 07 48 29 0 4 14 G8III 1 8 Figure 18 Suitable stars of type B G taken from the Gemini WWW site The Gemini MIR resources page also provides an automated script to suggest optimal spectroscopic calibration stars This 1s linked below and the screen shot below shows the input and output from the script Note that many of the standards are so called Cohen standards as they are drawn from Cohen s seminal work on standards 34 Version 1 2 2009 04 23 Target position 16 44 29 470 23 47 58 20 Nearest 1 HD152326 pos 16 51 45 26 24 39 23 2 d 1 8644 degrees COHEN 5 036 12 5 238 guidestar GSC0206201039 12 21 mag Nearest 2 HD149009 pos 16 31 13 43 22 11 43 6 d 3 4487 degrees COHEN 5 781 12 8 218 guidestar GSC0151801061 12 07 mag Nearest 3 HD156652 pos 17 17 34 65 28 54 47 8 d 9 0001 degrees COHEN 6 917 12 5 634 guidestar U1125 08135093 11 3 mag Nearest 4 HD158899 pos 17 30 44 31 26 06 38 3 d 10 7299 degrees COHEN 4 402 12 15 910 guidestar GSC0207900371 10 47 mag Nearest 5 HD143107 pos 15 57 35 25 26 52 40 4 d 11 0305 degrees COHEN 4 143 12 13 120 guidestar GSC0203701751 11 53 mag Nearest 6 HD142574 pos 15 54 34 61 20
11. This manual is presented in four main sections Section 1 gives an overview of the instrument Section 2 describes the detector Section 3 summarizes the optical information most relevant for users who are astronomers and Section 4 provides useful general information about mid IR astronomy with the attached appendices containing additional information specific to the operation and configuration of CanariCam In this version of the document certain assumptions are made regarding the system capabilities that are subject to change during commissioning Specific questions about the system should be directed to the GTC Help Desk 1 1 The Camera CanariCam is optimized for excellent image quality across the 8 26 um window The optical design is such that CanariCam 1s diffraction limited at 8 um Nyquist sampling the 8 um PSF The array is a 320x240 S1 As blocked impurity band BIB device with a pixel scale of 0 08 pixel providing a total field of view of 28 4 x 21 1 There are three powered mirrors and five folding flats four fixed and one movable on the grating turret All the mirrors are gold coated which provides a combined reflectivity of 92 CanariCam 1s equipped with a suite of narrow band and broad band 10 and 20 um filters located within two filter wheels immediately after the Lyot stop wheel at the first pupil image To reduce the cryogenic instrument size the optical layout is folded with the optics located on two sides of the optic
12. _ 0 10 0 00 7 00 8 00 9 00 10 00 11 00 12 00 13 00 14 00 Wavelength um Also located in the HWP wheel are the window imaging lenses which allow inspection of the window quality to confirm optimal throughput and minimize background The window imaging mode is an engineering mode and is not expected to be used by astronomers The full wheel 15 listed below Table 3 HWP Wheel Positions iret Open ______ Open for standard observations for standard observations HWP 10um HWP for 10um polarimetric observations e Version 1 2 2009 04 23 Open HWP1 First spare position for a future HWP Open Open for standard observations Open HWP2 Second spare position for a future HWP 2 4 Aperture Wheel The aperture wheel 1s located at the telescope focal plane and coincides with the entrance to the optics train housing This wheel contains a selection of two field stops a mask containing 3 pin holes a pinhole grid two polarimetric field masks two coronagraphic masks a slit dekker not needed and a blank disc Aperture positions are summarized below in Table 4 Table 4 Aperture Wheel Positions Field Mask 1 10 oversized field mask default mask Field Mask 2 20 oversized field mask Polarimetric mask for imaging and or spectro polarimetric mask 1 observations Polarimetric mask for imaging and or spectro polarimetric mask 2 observations Unused in current Canar
13. a housing that is continuously purged by dry gas The active window 15 also bathed in dry gas In the protected position an aluminum blank is used in place of a transmissive window so that none of the windows 1s exposed when the instrument 15 not in operation An external warm stepper motor rotates the wheel so that any of the windows can be rotated into the incident beam Table summarizes the properties of the window materials Figures 6 8 shows the transmission curves for each window as provided by the window vendor Janos Technology Also in each figure in the bottom panels are shown measured but with limited wavelength coverage transmission curves for CanariCam where there is overlap in the graphs the measured graphs should be adopted Table 1 Entrance Window Parameters Window Diameter Thickness KRS 5 48 mm ZnSe 50 mm BBAR 0 4 KBr Humidity 9 Block Safe Storage Position Block mirror Mirrored safe storage position ie Version 1 2 Thallium Bromoiodide 2009 04 23 60 100 100 D 80 vw 60 E Ww cC 5 F BR EN p i i B 02 04 06 08 1 0 2 0 40 6 0 20 30 40 Wavelength um KRS Transmission vs Wavelength 74 00 72 00 70 00 68 00 E Transmission E 66 00 5 64 00 62 00 60 00 0 00 2 00 4 00 6 00 8 00 10 00 12 00 14 00 16 00 18 00 Wavelength um Figure 8 Transmission curve for KRS 5 wi
14. are incorrect Wavelength calibration 15 accomplished using a sky spectrum as arc spectra are not available in the N nor Q band The nswavelength command is called which in turn calls the gnris RAF Gemini package which in turn calls the identify NOAO IRAF package The default N and Q band spectroscopic line IDs are in the data directory of the gnirs directory but the lines can actually be blends or band features in some cases such as in the case of the 9 503 um peak of the ozone feature commonly used to calibrated the low resolution N band ee ee Version 1 2 2009 04 23 spectra The extent to whish lines bands are distinct from sky emission is a strong function of the conditions and most especially the water vapor column along the line of sight as well as the slit width used N band A Q band X um 7 467 17 586 7 875 18 298 8 514 18 648 Nu Version 1 2 2009 04 23 Figure 20 N band spectra during the line ID phase using data taken in dry conditions with a narrow slit Above the N and Q band line listing Initially one should check for correct line ID but as said above the automatic line IDs are typically incorrect requiring the lines to be cleared and re identified using the 1 key Below are listed a sample line ID taken from the example above In the figure and table below we show a spectra taken in worse wetter conditions and with a wider slit The line identification 1s les
15. clear The other three positions contain spectrally flat sources that are currently being tested for use in creating flat fields for both imaging and spectroscopic modes Also in this wheel is the wire grid polarizer used for calibration of the polarimetric modes of cc Table 2 Sector Wheel Positions Sector Wheel M Comments Position Open hole normal operation 16 Version 1 2 2009 04 23 Polarizer for polarimetric calibration Black Plate High blackbody for flat fielding Polystyrene Increases background for flat fielding 2 3 Half Wave Plate Wheel To minimize any instrumental polarization the half wave plate HWP sometimes called the half wave retarder wheel 15 the first optical component after passage through the entrance window This wheel can accommodate up to three HWPs but currently only the 10 um HWP is installed In the future a second HWP may be purchased and installed possibly to cover the 20 um waveband or to provide achromatic polarimetric retardation versus wavelength The HWPs can be rotated to four position angles 0 22 5 45 and 67 5 The retardance efficiency versus wavelength and the laboratory transmission of the combined HWP and Wollaston was measured to be 83 at 10 5um Polarisation Efficiency Retardance 1 00 4 0 90 0 80 0 70 0 60 0 50 0 40 Retardance Efficency 0 30 0 20 e 10um HWP Retardance 10um HWP Efficiency
16. confirmation the angles of 45 and 67 5 are typically used for more on polarimetry basics the reader 15 referred to the polpack manual http star www rl ac uk star dvi sun223 htx sun223 html or any of the many discussions of polarimetry data collection in the literature The Wollaston prism is the analyzer which splits the incident beam into two beams that are perpendicularly polarized The beam that passes through the prism that is undeviated 1s called the ordinary beam or o ray whereas the deviated beam 1s called the extraordinary beam or e ray Both beams are recorded simultaneously on the array and hence for each object or patch of sky in the field of view two perpendicularly polarized images are formed on the array For this reason it 1s typical to use the polarimetric mask in the aperture wheel to obscure 50 of the field of view This prevents overlapping of the images or crosstalk from residual sky emission As CanariCam uses two transmissive components the HWP and Wollaston prism there are two key chromatic effects The separation of the o and e beams is a function of wavelength as shown in figure 31 below from lab measurements This change in dispersion versus wavelength 15 an interesting effect for imaging and one should note the change when performing the polarimetric data reduction In spectropolarimetry mode the resultant spectra will be curved and again this must be noted when extracting the spectrum Disper
17. for example telescope tracking errors differential refraction and flexure This 1s simply and readily achieved with the use of the flip mirror located on the Wollaston prism slide see above The procedure to confirm the source position with respect to the slit 15 to set the instrument for standard spectroscopic observations and then use the flip mirror to divert the beam before it reaches the grating Although the resultant image 1s somewhat vignetted and aberrated the optical path with flip mirror inserted deviates from the optimum path the source location is still indicated adequately Once this confirmation has taken place the flip mirror is retracted and normal observations continued Some non repeatability in the positioning of the grating turret can occur so one should confirm the position of the grating with care before starting observations In the low resolution modes the dispersion in the N and Q windows is such that light will never be dispersed beyond the array even if there is a slight error in the turret position which however may vary from night to night This does not pose a problem for the data reduction but one should still be aware of this possibility The positional repeatability of the gratings and main imaging mirror can be seen in Table 10 which shows the mean one sigma error Table 9 Grating Wheel Positions Grating Wheel Position Description Mirror position for imaging LowRes 10 10 um low resoluti
18. handle these high data rates From an astronomical perspective these high background fluxes have a more serious consequence than just the necessity of fast acquisition electronics astronomical objects of interest are invariably much fainter than the background emission Thus very precise subtraction of this background flux is required to extract the signal of interest For exam ple typical bright infrared standard stars used for flux calibration are frequently an order of magnitude fainter than the background emission And scientifically interesting astronomical objects may be fainter than the background emission by four or more orders of magnitude 1 1 The Standard Chop Nod Technique The requirement of precise background subtraction dictates the method by which images are acquired at a telescope Background subtraction is effected in real time using the standard infrared astronomical chop nod technique In this technique the telescope 1s pointed at an object of interest the program object and the camera acquires a set of images An image consists of signal from the program object added onto the much larger signal from the background The secondary mirror of the telescope is then moved slightly away from the nominal position so that the program object moves out of the field of view of the camera and another set of images is acquired This procedure called a chop cycle is repeated many times at typically a 2 10 Hz rate movin
19. likely gain the most from GTC s low emissivity and the outstanding Observatorio del Roque de los Muchachos ORM site characteristics This advantage 1s dramatically apparent when one realizes that in the background limited regime that applies to the mid IR region the integration time required to achieve a given signal to noise ratio for a point source scales inversely as the telescope aperture diameter to the fourth power Thus remarkably an observation takes 50 times less integration on the GTC than at a 4 meter class telescope even if the difference in emissivity is ignored These gains imply that previously impossible scientific programs can become routine with the GTC The CanariCam design is driven by the science that takes full advantage of these opportunities ill Version 1 2 2009 04 23 TABLE OF CONTENTS 1 J str ment IntroduetioB s ccs opc n Den THU Man ue EN M DM MUN 1 M ere ey eee me ears ease ne eee 1 1 2 SD6c lt OPPIDI 2 Bo Goode tere dedi boe dates hores attesa ned 2 1 4 The Mi esea T A RO 2 kS Basic Instrument C Daractetts CS 4 LO Deer Reddot ATE 6 L 6 l Simele Pomn Sampung 91 sera te A AE 6 L62 Quadruple sampling SLR Janean agi att 7 1 7 Mid Infrared Plat etait oa de eee etti iste xod a dos deed 9 LS Detector Temperature C Ono
20. of readout direction which is up on the displayed image e Pattern 2 negative offset in the same column as the peak of a bright source Pattern 3 a negative offset in the same row as the peak of a bright source e Pattern 4 a horizontal see displayed image sequence of negative spots along the entire row where the bright source 15 positioned Figure 3 which is courtesy of the COMICS team shows all four effects The negative offset or level drop Patterns 1 and 2 are due to the transient variations of the characteristics of the source follower circuits used to read out the array when the input signals are drastically changed as occurs during read out of a bright source The negative offset or level drop Patterns 3 and 4 are due to the variations of the bias voltages when the input signals are changed due to a bright source The CanariCam teams expresses its deep thanks to the COMICS team who have kindly provided numerous materials about the Raytheon array and their improved read out technique domo arigato gozaimasu Version 1 2 2009 04 23 Figure 3 Image of bright star taken with the instrument COMICS on Subaru Four negative offset or level drop effects are shown in these images as described in the text 1 6 2 Correlated Quadruple Sampling S1R3 Effects 1 and 2 can be eliminated by using a readout method called Correlated Quadruple Sampling CQS which consists of referencing the detector
21. ramp voltage to several of the silicon multiplexor or mux internal voltages In order to reduce Patterns 3 and 4 which are seen only when observing very bright objects analytical corrections are required in addition to CQS These analytical corrections are not generally available as the effects are a function of the frametime source brightness and other variables Though noticeable this effect is relatively small in CanariCam resulting in only a 0 1 0 6 effect in signal as determined from the ratio of the source peak signal to that of the negative offset The exact level of that offset 1s dependant on the frame time and source intensity but not the distance from the source see below for a detailed discussion based on our laboratory tests It appears that this artifact may be partially eliminated by referring to channels that show the spots but contain no source Although these methods effectively eliminate or greatly reduce readout artifacts associated with the Raytheon CRC 774 detector array they also introduce new operating limitations as follows e Since COS requires four times the data to correct for patterns 1 and 2 frametimes must be approximately four times longer Thus the minimum frametime for CanariCam of 5 ms is effectively increased to 20 ms While not limiting most observations with CanariCam it does hamper observations using broad band filters such as N band and Qw 20 8um These wide pass band filters ex
22. the 4 frames contains an image of the source with signal S and noise Ofame AS described earlier in Section ILII and as shown in Figure 37 the two chop differenced frames will each contain two images of the source one positive and one negative The individual S N of each of these source images 1s Si V2 Oframe This 15 of course the same as for the single source image in each of the two chop differenced frames as derived for the standard method Now the net signal is formed by taking the difference of the two chop differenced frames The net signal frame contains four images of the source two positive and two negative The key point here is that in forming this net signal difference the positive and negative images in a given chop differenced frame are each combined with a portion of the other chop differenced frame which does not contain signal recall in the standard method that the single source image in each chop differenced frame 15 located at the same position Therefore the S N of each of the four source images in the net signal frame is reduced by V2 with respect to its value in the chop differenced frames This yields a S N of S 2 Oframe per source image The final image is formed by cutting the net signal frame into four pieces each containing a single source image registering them and summing them In this process the source signal increases by a factor of 4 and the noise taken in quadrature increases by i e JA The fin
23. to permit adequate correction for radiative offset see appendix below for a detailed explanation of the reasons for chopping and nodding The exposure time can be a confusing term when applied to mid IR observations there are four exposure time commonly expressed 1 Frame time This is the time between readouts of the array and is controlled through the optimized software settings Typical times are 25 ms 2 Saveset time This is the time between each co added dataset that 1s saved to disc This is the smallest quantum of data saved by CanariCam This is controlled through the optimized software settings Typical times are 10 s 3 Exposure time This is the user selected time interval over which the source will be observed The time can vary greatly depending on source flux but typically it is in the range 60 600s 4 Clock or total time This is the total time need for the observation to be completed and includes the necessary overheard of chopping and nodding The multiplicative factor to convert exposure time to clock time 15 dependant on the observing mode chop and other parameters used but is estimated to be 3 for imaging 4 for spectroscopy and 3 75 for imaging polarimetry Thus a typical dataset may consist of 25 ms frames of 10 s savesets with a 600 s exposure time which results in a clock time of 1800 s It 1s important to note that only the exposure time is input by the user All other parameters are set automaticall
24. 6 A zoomed in image of the fourier spectrum where the white horizontal line shows the filtered real and imaginary parts of the spectrum After exiting from the fourier domain plot as usual in IRAF using the q key the image below 15 displayed showing the original and de fringed spectra If the de fringing was successful the program is exited with the r key or to exit and reject the de fringed spectra this is achieved with the 1 key 4 4 7 Spectral Flux Calibration Spectral flux calibration is achieved with the msabsflux task This in turn calls the telluric task in the IRAF NOAO onedspec package In most cases it is wise to carry out this step interactively since the automatically found best spectrum often 1s not actually what is wanted especially for low resolution spectra typically due to having large regions of little or no signal in the source and calibration spectra Noise in these regions can often produce absurdly large signals in the resultant spectra which severely skews the chi squared minimization used in the IRAF telluric task to find the optimum shift between the two spectra In view of this problem the first thing one should do in running the task interactively 1s to window the plot to exclude the long and short wavelength regions with little or no signal This is typically the wavelengths gt 13 um and lt 7 5 um One way to proceed is to set offset 0 and dscale 0 so only one output spec
25. 8 190 guidestar GSC0260401448 12 94 mag Figure 19 Automated standard star selection too taken from the Gemini WWW site Many Cohen spectrophotometric standards are early K dwarfs and many have accurate IRAS mid IR flux densities making them potentially good flux calibrators as well as telluric line removal stars At low spectral resolution the fundamental vibration rotation band of SiO significantly depresses the spectrum at 7 5 10um in stars later than around K2III This can significantly affect the ratio ing but the effects can be removed or mitigated through use of a Cohen model template or through the Gemini IRAF msabsflux Any Cohen standard star can be used for both low spectral resolution N and Q band spectra However science programs that are aimed at detecting or measuring detailed shapes of weak spectral features in a strong continuum source in the N or Q band should use Cohen stars with care At high spectral resolutions the Cohen standards are unsuitable as telluric standards as they contain a multitude of photospheric absorption lines Finally we note that another set of Cohen standards termed tertiary Cohen standards are available but these have not been well characterized observationally and so their quality as standards is generally unknown and hence primary or secondary Cohen standards are recommended instead 35 Version 1 2 2009 04 23 4 4 1 IRAF Reduction of Spectra The spectroscopic eq
26. al bench The two sides are connected optically only through a pair of folding flats F3 and F4 with the bench itself serving as a very affective baffle reducing stray light that can reach the focal plane In addition to the normal field imaging mode CanariCam is capable of both pupil and window imaging Pupil imaging is achieved by re imaging the first pupil image onto the focal plane with a ZnSe relay lens The pupil imaging mode is used to ensure that CanariCam is optimally aligned with the telescope optical axis This diagnostic capability is particularly important since CanariCam can be mounted at any one of several different focal stations Nasmyth the up looking straight cassegrain port or the side looking folded cassegrain port each of which will have slightly different alignment constraints The window imaging mode permits one to monitor the health and status of the dewar windows Several windows are mounted on a plate attached to the dewar with a ferrofluidic feedthrough so that any one of the windows can be inserted into the beam during normal operation This window changer assembly thus permits one to optimize the window material for specific ambient observing conditions and scientific observing requirements Window imaging allows the observatory to maintain a record of the quality of each of the windows to determine if a particular window may need replacing Version 1 2 2009 04 23 1 2 The Spectrograph CanariCam is designe
27. al source S N becomes 4 S 2 2 Ogame S Ogame This is the same S N as obtained with the standard method Note well that this only holds for the background photon limited noise regime M S
28. and A 4 at 10 um This mask is painted black and 15 attached to the center of a thin ZnSe disk that 15 inserted into the focal plane mask wheel The pupil mask consists of a small central opaque disk which blocks the central obscuration in the pupil image and a large hexagonally shaped mask that blocks the outer edges of the primary mirror The central opaque disk of the mask 15 140 the size of the image it is masking and the outer hexagonal mask 15 90 the size of the edge of the pupil image it is masking The disk shaped central part of the pupil mask is supported by thin slightly wedged veins that mask the images of the secondary mirror spider supports With the occulting and pupil masks in place the total throughput of the coronagraphic mode 15 about 66 that of the regular imaging mode At the beginning of an observing sequence one must have the bright star centered on the occulting spot and the pupil mask rotated so that the spiders are masked see below The pupil mask is mounted in a holder that rotates in coordination with the rotation of the pupil which occurs as the alt az telescope tracks an astronomical source across the sky However the motor that rotates the mask cannot be turned on during data acquisition which would introduce noise so pupil mask rotation can only be permitted while the telescope is beam switching nodding to its alternative position Itis for this reason that the spider mask is somewhat oversized 20 t
29. aration and data reduction We explain the methods for taking raw data and producing a flux and wavelength calibrated spectrum As with the imaging data above the Gemini mid IR IRAF package is key to accomplishing this goal and we are particularly indebted to the Gemini staff who wrote the MIR resources Gemini WWW pages especially Kevin Volk Rachel Mason and 38 Version 1 2 2009 04 23 Scott Fisher The section will concentrate on the msreduce package which can be used with The single spectrum of an object and resulting in a wavelength calibrated spectrum Two spectra source and telluric standard and resulting in a wavelength and flux calibrated spectrum Implicit in the above is that two spectra are observed one for the object and the second for calibration Flat fields are often taken but are unimportant for compact sources For the calibration source a star or asteroid are often used Asteroids are preferable when using higher spectral resolution as some stars can show resolvable spectral features At lower spectral resolution stars are often better since one can use these objects as both flux and wavelength calibrators Telluric standards are used to remove effects of atmospheric transmission Based on experience on Cero Pachon and Mauna Kea for objects observations lasting gt 30 minutes elapsed time two telluric standards are suggested post and prior to observing In the case of Gemini one of those is free
30. arger than projected pupil size 2 6 Filter Wheels CanariCam has two filter wheels located just downstream from the Lyot wheel There are 13 positions in each filter wheel for a total of 26 available positions Currently six positions are either open or used for engineering tasks including the pupil imaging lenses used to optimize pupil alignment of the instrument to the GTC The other 20 positions contain 1 inch diameter circular filters for astronomical use Table 6 presents a summary of the filter characteristics The filter curves are shown in the GTC s CanariCam WWW page In addition there is a very slight but noticeable difference among some of the filters in the optical beam direction after passage through the filters This results in a slight shift in the location of an astronomical image as one changes filters Table 6 Summary of Filters Filter Central Bandwidth Name um Comments Further information on the specific filter characteristics is available from the CanariCam WWW page or in the CanariCam Optical Design Manual Ox Version 1 2 2009 04 23 WU m Name um um EE EGER er ou 04 789 1 00000 __ 53 98 10 arja Sa 13 09 oe 3 05 sa 3 5 0 Sic Las D fafo p oe Nem Ds 03 0000000 09 __ 5 10 j 5 j Part of VISIR co
31. cation of the XREGISTER task of IRAF to shift the images to a common peak of a source before co addition or co averaging This task requires a bright point source in each of the savesets for reliable operation and hence the MISTACK task is more commonly used than MIREGISTER X xterm Image Reduction and Analysis Facility PACKAGE midir TASK miregister inimages Input T ReCS or Michelle image s outimag Output image s outpref x Prefix for output image s raupath Path for input raw images combine average Combining images by averagelsum fl vari no Output variance frame regions Regions to be used for registration logfile Logfile verbose yes Verbose status 0 Exit error status 0 9009 gt 0 bad scanfil Internal use only mode ql fo 2 Figure 17 EPAR of MIREGISTER 4 3 5 MIREGISTER This task calls most of the previous packages in a non interactive manner and 15 the easiest most automated and most commonly used manner to reduce data The task calls TPREPARE MISTACK or MIREGISTER and can call TVIEW and TBACKGROUND if those flags are set to true The output file is a 320x240 pixel file in the 1 extension whose 0 extension is the full FITSHEADER Once the data is in this format manipulation and measurements can be performed using standard FITS tools and or packages 44 CanariCam Spectroscopy Data Reduction In this section we discuss mid IR spectroscopy prep
32. d on accurate photometry that the flux calibrator be observed close in time and nearby on the sky to the science target Version 1 2 2009 04 23 G Mc infrared information Mid infrared Standard Star Brigh 3 6 id infrared information http staff ge ss one amp mag Fnu C 6 information L http staff ge ne amp mag Flambda E In Band Brightness Values for Mid Infrared Standard Stars Estimated fluxes for alpha_cma Estimated fluxes for alpha cma This page allows one to find the estimated in band fluxes of standard stars for MICHELLE or T ReCS The values are T ReCS brightness values T ReCS brightness values calculated from the spectral energy distributions and the filter profiles and include the effects of the blocking elements They do not allow for the mirror reflectivity and detector quantum efficiency as functions of wavelength E E 2l N Broad 10um 129 66901 Jansky N Broad 10um 4 03398E 12 Watt meter square micron Calculations were carried out for the assumed outside the atmosphere case for a standard zenith atmospheric transmission and Q Broad 20 8um 32 68500 Jansky Q Broad 20 8um 2 27014E 13 for the atmospehric transmission at 2 airmasses These values together give some indication of how sensitive the filter is to Qshort 17 65um 42 87000 Jansky Qshort 17 65um 4 08658E 13 atmospheric extinction However it must be borne in mind that the extinction is a function of the water vapour col
33. d to be a low to moderate resolution spectrograph at 10 and 20 um Four gold coated plane diffraction gratings are located on a rotating grating turret In spectroscopic mode the gratings replace the folding flat located on the turret The two low resolution gratings cover the entire 10 um window 175 AA 7 5 13 5 um and the entire 20 um window R 120 AX 16 0 25 0 um The 10 um moderate resolution grating R 1300 spans AX 0 65 um when centered at 10 5 um and requires several settings dependant on the amount of wavelength overlapping required to cover the entire 10 um atmospheric window The 20 um moderate resolution grating R 890 spans AX 0 45 um when centered at 20 5 um and also requires several settings dependant on the amount of wavelength overlapping required to cover the entire 10 um atmospheric window The spectrum 1s dispersed along the long axis of the array 320 pixels For each of the low resolution grating settings the spectral resolution element is sampled by approximately five pixels The slit lengths are oversized to ensure that the slits span the entire short dimension of the array Nine slits are available offering flexibility in matching the wavelength of operation or the seeing conditions 1 3 The Polarimeter CanariCam uses a CdSe Wollaston prism and half wave retarder half wave plate HWP to provide dual beam polarimetry at all wavelengths in the N band window The peak accuracy of the instrument
34. determined during CanariCam commissioning on sky The result is that for a 100 polarized source measured at 10um CanariCam measures 100 polarization whereas for a 100 polarized source measured at 8um CanariCam measures 80 polarization One must note these values are for monochromatic light and so for both broad and narrow band filters centered even close to 10um a polarimetric efficiency correction must be made As the response is of the HWP efficiency vs wavelength has been well characterized in the lab a simple multiplicative correction 1s required to report the true degree of polarization the position angle PA of polarization in unaltered by the HWP efficiency _5 Version 1 2 2009 04 23 Retardance Efficency 10 0um HWP 9 Retardance m Efficency 7 00 8 00 9 00 10 00 11 00 12 00 13 00 14 00 Wavelength um Figure 33 Retardance and efficiency of the CanariCam 10um HWP as a function of wavelength The essential data reduction for the data is described in the following section Data 15 collected using the standard chop nod procedure of MIR observations but we note that currently it is unclear how many HWP PAs will be required to give an optimal polarimetric signal per nod position This is an essential task for commissioning of the CanariCam polarimetric mode It is crucial to note that the images are not stacked across HWP PA s if this is done the polarimetric data i
35. duction of TPREPARE most tasks do not require its use and hence we do not consider this task further 4 3 3 TVIEW This task is used to display the savesets to the display of choice e g ds9 to allow investigation of the images This can be done automatically where the images are automatically displayed to the image tool or interactively where image parameterization can be performed using IMEXAM After examination in interactive mode nods can be flagged as bad and excluded in subsequent data reduction As there must be an even number of nod positions TVIEW ensures the output file achieves this 30 Version 1 2 2009 04 23 X xterm Image Reduction and Analysis Facility PACKAGE midir TASK tview inimages 52004020150118 fits Input T ReCS image s outimag Output image s outpref v Prefix for output image name s rawpath Path for input raw images type dif Type of frame srclrefldif delay 0 Update delay in seconds Fl inte yes Interactive screening of frames fl disp no Fill display fl labe yes Label display colour black Label colour blackluwhitelredlareenlblueluellow fl use no Use imexam to look at images before screening l sh c no Show changes to image headers 1 dele yes Delete unchanged copy images 21 0 Minimum level to be displayed z2 0 Maximum level to be displayed zscale yes Display graylevels near the median zrange yes Display full image i
36. e array and correspondingly large variations of detector power dissipation Due to the finite response of the closed loop temperature control system it may typically take a few seconds for the detector temperature to settle to the control temperature to within the stability requirement No observation should begin until the temperature reaches the required level The CanariCam Instrument Sequencer will be designed to enforce this unless manually overridden 10 Version 1 2 2009 04 23 2 Instrument Optics Here we describe the function and key properties of each component in the optical train beginning at the entrance window and proceeding along the light path to the detector All mirrors are gold coated diamond turned aluminum and all internal wheels are moved with cold stepper motors Many mechanisms such as the Aperture Pupil and Lyot Wheels contain components that will usually be inserted by GTC staff at the beginning of an observing run and remain unchanged during the run Figure 7 shows the optical design for CanariCam ell Version 1 2 2009 04 23 l Window Imaging Window Lens 2 Filters amp Pupil Imaging Lenses HWP Window Imaging Lens 1 Wollaston AS NN Wh Array Prism Son F5 P2 Figure 7 Sketch of the optical layout of CanariCam The optical train on the upper deck of the optical bench is at left and the optical train on the lower deck 15 at right Optics after the folding mirror are locat
37. e as on source minus off source the source signal remains positive in both chop differenced frames The signal levels in these differenced frames range 3 2x10 electrons 5 which is 0 4 of the raw background signal Finally the bottom row shows the net signal obtained by adding together the two chop differenced frames shown in the middle row note that no other processing has been done to the data other than the additions and subtractions as described above The detected source is the nuclear region of the starburst galaxy NGC 253 The net signal is the result of a total exposure of 20 minutes in which half that time 1s actually spent imaging the off source reference position The 58 _ Version 1 2 2009 04 23 signal level at the tail of this source near the middle of the frame is 6 4x10 electrons s This is about four orders of magnitude below the background level shown in the raw frame In fact the signal to noise ratio at this level in each pixel 1s about seven so that the effective background subtraction 1s more nearly five orders of magnitude below the background Nod Position A Nod Position B On source OFF source On source Off source src 1 ref1 src2 ref2 On source Off source On source Off source dif1 dif2 Net Signal Figure 36 Illustration of standard chop nod technique Common terminology or short hand for the individual frames is given in brac
38. eam at any A within the N band CanariCam uses a Raytheon CRC 774 320 x 240 S1 As blocked impurity band BIB or impurity band conduction IBC high background infrared Focal Plane Array FPA The detector has excellent cosmetic quality Figure 1 shows an image of the detector under uniform illumination in standard chop mode There are no dead pixels Some of the key detector characteristics are Detector Raytheon CRC 774 320 x 240 Si As IBC Physical Pixel Size 50 x 50 um Full well Capacity 1 0e7 or 3 0e7 electrons high or low gain Array Readnoise Deep well 3 6 ADU 2060 Shallow well 4 6 ADU or 828 Dark Current lt 100 electrons sec T 6K Quantum Yield gt 40 at peak Number of Output Channels 16 Version 1 2 2009 04 23 Figure 1 Dark cold internal blankoff chopped image see Appendix A Section ILI taken with CanariCam Raytheon CRC 774 320x240 100 4 50 40 30 20 fe 5 gt a SSeS SSS 5 A No Coating Approx Limit w no AR coating Wavelength Figure 2 Quantum yield percent of a typical Raytheon CRC 774 320 x 240 Si As IBC detector Version 1 2 2009 04 23 1 6 Detector Readout several mid IR instruments on large telescopes use the Raytheon CRC 774 320 x 240 Si As IBC detector array These include COMICS on the Subaru 8 m Kataza et al 2000 Michelle on the Gemini North 8 m Glasse et al 1997 the new dec
39. ed on the lower deck of the optical bench 12 Version 1 2 2009 04 23 2 1 Dewar Entrance Window Maintaining a high quality entrance window is essential for CanariCam to routinely achieve its performance requirements Windows must be clean dry and largely free of scratches to reduce scattering of light and to avoid significantly increasing the thermal emission that would reduce frame times and increase the background photon noise Furthermore since CanariCam must remain at low temperature and on the telescope for months at a time the entrance window assembly must be robust enough to ensure that the windows are well protected from dust and moisture for long periods of time and under sometimes harsh conditions Degradation of windows by water is a particular concern since some of the window materials are water soluble To address these issues and concerns CanariCam employs a window turret design in which the cryostat window changer 1s used to select the optimum window for the selected instrument configuration wavelength of observation and environmental conditions in particular humidity It utilizes one stepper motor and has the following five positions ZnSe window KBr window in two of the positions KRS 5 window and two blanks one is mirrored and gold coated to minimize background radiation during long periods when CanariCam 15 not in active use Only the active window is physically exposed all the others being protected by
40. efereed to as atmospheric windows In the mid IR there are two major windows The first is located between 8 14 um while the second is located between 16 30 um often called the 10 and 20 um windows respectively Within these mid infrared windows there can be rapid variations in transmission due to changes in the water vapor column depth This can be seen in the change of transmission in Figure 35 when the atmospheric water vapor changes from 1 0 mm to 3 0 mm In addition to absorption the atmosphere also emits strongly in the mid IR peaking at 10um Thus the atmosphere not only attenuates the mid IR signal from an astronomical source but also dilutes the signal with thermal emission of its own Separating the background sky emission from the source emission 1s one of the key elements in observational mid IR astronomy Transmission Wavelength um Figure 35 Atmospheric transmission at Mauna Kea Hawaii The 7 14 micron window can be seen clearly Longer wavelength filters take advantage of the 16 30 micron window which has numerous absorption features Image adapted from ATRAN model Lord S D 1992 and data from Gemini Observatory 56 Version 1 2 2009 04 23 Another affect the atmosphere has on observations is seeing This is a blurring affect caused by turbulence density inhomogeneities 1n the atmosphere This results in random fluctuations in refraction causing a star to vary in intensity and location on
41. es a noise spike at 1 Hz Thus the chop frequency 15 a carefully selected value based on both the array characteristics and typical sky conditions 4 3 CanariCam Data Structure amp Exposure Times CanariCam data is delivered in standard multi extension FITS MEF file format Each extension contains the 320x240 pixel image as well as specific headers relevant to those extensions The zeroth header is a general header containing greater information about the full data file When CanariCam is setup for engineering modes or unusual observing modes the instrument GTC may not use chopping and thus would not nod In this case the images are stored as 320 240 1 where N is the number of savesets Savesets are automatically defined by the CanariCam software to optimize for observing efficiency and full well depth and can be adjusted for specific observing routines but this 1s rarely needed typical imaging frame time 15 25 ms and a saveset time 15 10 s In chop and chop nod modes there are 2 chop positions per saveset and M savesets meaning that each image is therefore 320 240 2 M Almost all science data are taken in 25 Version 1 2 2009 04 23 chop nod mode and each extension contains the savesets for a single nod position CanariCam can be used in the nod sequences ABAB or ABBA with the difference being only an insignificant change in overheads However it 1s essential for a minimum number of AB pairs to be observed
42. escope slew as the science object to minimize the change of gravity vector pupil rotation and temporal changes of the PSF There are relatively few flux standards available for use with 8 10 m class telescopes since many standards saturate when observed with these large apertures Flux standards are typically drawn from one of the following lists e Primary or secondary Cohen standards e Very bright southern standard stars but many of these sources may saturate CanariCam GTC unless there is a telescope defocus The Cohen standards are developed from models of continuous spectra of many stars tied carefully to observational data These model spectra can be used for imaging and spectroscopic flux calibration by integrating over the filter bandpass or smoothing to the appropriate spectral resolution A detailed discussion of this work can be found in Cohen et al 1999 AJ 117 1864 but we note here that the calibration is anchored to two primary standard Alpha Lyr A0 V and Alph CMa Al V Gemini has some outstanding tools to assist with this procedure see http staff gemini edu kvolk brightness html which calculates the estimated fluxes for the filters fitted in T ReCS and Michelle We note that airmass corrections at mid IR wavelengths are not usually done unless large blocks of time are available for a specific program to permit observations over a broad range of airmasses It is strongly recommended for programs that critically depen
43. etry are discussed as part of the polarimetry data reduction section at the end of this document Table 8 Wollaston Slide Positions Position Description Wollaston and mirror out of beam for all standard Clear observing modes Flip mirror in the beam see grating turret below Wollaston Wollaston in the beam for polarimetric observations 2 9 Mirror amp Grating Turret The grating turret provides the main mirror for all science imaging and four gratings for the spectroscopy modes The turret utilizes a single stepper motor which may rotate into the science beam any of the following five components imaging mirror low resolution 10 um grating high resolution 10 um grating low resolution 20 um grating high 2 Version 1 2 2009 04 23 resolution 20 um grating Table 9 When using the high resolution gratings the grating may be positioned to center any wavelength within the operating range on the detector array Multiple settings of the high resolution gratings are needed to span the entire 10 or 20 um atmospheric window When using either of the low resolution gratings a single setting places the entire operational wavelength range for that grating on the detector array The gratings are optimized for use in first order and are gold coated master gratings on aluminum substrates While using CanariCam in the spectroscopic mode it may be desirable to monitor the source location in the slit the source position may drift due to
44. g back and forth between yee Version 1 2 2009 04 23 on source and off source positions chop differenced signal is formed by taking the difference between the on source and off source images While this rapid movement of the secondary mirror allows subtraction of a spatially uniform background that is varying in time at frequencies below the chop frequency it usually still contains a spurious signal that may still be significantly larger than the source signal This signal termed the radiative offset results from the fact that the emission pattern of the telescope as seen by the camera depends on the optical configuration of the telescope Movement of the secondary mirror changes this configuration resulting in two different emission patterns The difference in these emission patterns 1s found in the chop differenced signal In order to remove the radiative offset the entire telescope 1s moved after a short period of time so that the source now appears in what was previously the off source position of the secondary mirror This movement of the telescope is termed a beam switch or nod and typically occurs on timescales of tens of seconds Timescales are much longer than that used in chopping due to slower changes in telescope emission relative to sky emission Chop differenced frames are then formed with this new on source and off source configuration In this new configuration the radiative offset will have chan
45. ged sign and is effectively canceled when the new chop differenced data is added to the old chop differenced data provided the telescope emission has not changed in the time between beam switching Figure 36 demonstrates the acquisition of data in the standard chop nod mode The images shown here were obtained with University of Florida mid infrared imager spectrometer OSCIR at the NASA Infrared Telescope Facility in Hawaii IRTF The top row of four 1mages shows the raw data frames from the two secondary mirror positions at each of the two nod positions called Nod Position A and Nod Position B of the telescope These images are dominated by fixed pattern offsets due to pixel to pixel variations and offsets between the 16 channels of the acquisition electronics The background counts in these raw images correspond to 7 4x10 electrons 5 Each raw image consists of 5 minutes of total integration time i e 15 000 frames coadded using 20 ms frame integration time and the N band filter obtained in the chop nod sequence as described above The second row of two images shows the chop differenced data derived from the subtraction of the on source and off source data in the two nod positions of the telescope Note that the dominant pattern principally a gradient along the diagonal connecting the lower left to upper right corners of the images changes sign between the two chop differenced frames However since the subtraction 1s always don
46. gt 99 at 10 5um but rolls off significantly toward the cut on off of the N band window Correction for this will investigated as part of the commissioning process on the GTC and described at a later date To eliminate overlapping of the orthogonally polarized beams on the array a focal plane mask 1s used in the aperture wheel All parameters for CanariCam polarimetry are the same as those for imaging but around 50 of the field of view is blocked by the focal plane mask Also observing efficiency is slightly lower due to the repetitive motion of the HWP Nevertheless CanariCam polarimetry represents the first dual beam MIR polarimeter and hence opens the door to a near era of high precision MIR polarimetry 1 4 The Coronagraph CanariCam has the first mid IR coronagraphic mode ever used at a major ground based observatory This mode improves the ability to detect faint mid IR point sources such as sub stellar objects and faint extended sources such as circumstellar disks that are located very close to bright point sources and which might not be detectable without the coronagraphic mode To implement this mode one inserts an occulting spot mask at the telescope focal plane and a pupil mask at the first pupil plane inside the camera In the initial implementation E Version 1 2 2009 04 23 of this mode the occulting spot is sharp edged top hat opaque metal disk with a radius of 0 84 arcsec which corresponds to A 5 at 8 um
47. iCam configuration Coronagraphic Coronagraphic spot to be characterized during commissioning spot alpha Coronagraphic Coronagraphic spot to be characterized during commissioning spot alpha Alignment spots 3 pinhole spots 400 100 and 400um diameter used for alignment and image quality check out Alignment grid Grid of pinhole spots 100um diameter used for alignment and image quality check out Slit dekker Can be used in spectroscopic mode to reduce background light unlikely to be needed during standard observations to be confirmed during commissioning Blank to stop light passing past the aperture wheel 18 Version 1 2 2009 04 23 2 5 Lyot Wheel The Lyot stop is a crucial thermal baffle in the CanariCam optical system The Lyot wheel contains 4 positions at which are located the masks listed in Table 5 The Lyot wheel contains pupil masks for optimal thorough put and or rejection of scattered and or diffracted light Generally observers do not need to configure the Lyot wheel since it 1s configured automatically by the observing system or the support astronomer Table 5 Lyot Wheel Positions Circular Mask 1 Inscribed to maximum filled pupil Rose Petal Lyot Lyot mask most likely used only in coronographic mode The 6 segments as defined by the GTC s spider are reminiscent of a rose petal Circular Mask 2 Circumscribed to maximum partially filled pupil Circular aperture for science use 8 l
48. imes the actual image and slightly wedge shaped At the time of this writing the final procedure for centering a star on the occulting spot has not been fully determined The final procedure will depend on detailed telescope performance and will be revised during on sky commissioning of CanariCam on the GTC However the following procedure 1s one likely scenario for setup During normal operation of the coronagraphic mode with both masks inserted into the beam one centers the stellar image behind the occulting spot by moving the telescope first in right ascension centering in time then in declination also centering in time By centering in time we mean measuring the time at very slow slew to move the stellar image from one side to the other side of the occulting mask 1 e from disappearance behind the mask to reappearance Final adjustment of the stellar location is made by switching to the pupil imaging mode and making fine motions of the telescope that minimize the stellar flux in quick look frames While still in the pupil imaging mode the pupil mask is then rotated until the observer sees that the spider veins which emit mid IR radiation are blocked by the veins of the mask The spider mask should initially be positioned so that the spider image 1s blocked by but at one side of the mask During data acquisition the mask will not move but the pupil image will rotate during the telescope beam switch the motor is t
49. ing can be found for the mid IR instrument TIMMI2 Reimann et al 1998 see online manual Below is a short discussion of flat fielding in the mid IR as it applies to CanariCam In the optical flat fields are typically taken at the end of the night by observing either a screen mounted on the dome or the sky to provide uniform illumination on the detector However in the mid IR dome flats are unusable because the screen which 1s a room temperature blackbody will generally saturate a mid IR detector through most passbands In order to create a mid IR flat field at least two images with levels of background illumination different from each other by a few percent are needed Subtracting the lower background image from the higher background image provides a map of the spatial variations in sensitivity of the pixels across the array One method by which this can be achieved is to observe the sky at low airmass and at a higher airmass The Version 1 2 2009 04 23 background level increases with increasing zenith distance and atmospheric optical depth which results in the images being taken at both a low and a high background Another method permits taking the flat field near in time and location to the science observation This 15 done for CanariCam by use of the Sector Wheel Section 2 2 which contains a polystyrene mask This mask can be rotated into the beam to increase the thermal background and provide a high background image However this meth
50. ion and the PA of the vector 15 representative of the PA of polarization Software tools exist to take the I Q and U Stokes images and produce those vectors 7 Finally measurements of the images and vectors can be performed using standard image display tools There are few polarimetry data reduction tools generally available to the astronomical community with the notable exception of the Starlink program polpack This program is the most widely available and also most user friendly program the authors are aware of and hence we recommend use of this program to reduce CanariCam polarimetry data To run this program the Starlink software suite should be installed available free of charge from http starlink jach hawaii edu starlink We suggest the whole standard package is installed which will include the following key packages e Polpack e e Convert 4 5 2 Data Reduction Steps The data should be reduced using IRAF to the stage where there are multiple chop nod corrected files taken at a variety of HWP PAs Do not stack the images Should one wish to check the HWP PA this is recorded in the fitsheader The software that will automatically make the chop nod correction outputting a series of un stacked files at their respective HWP PAs will be produced after commissioning but this task can be trivially performed using the wfits IRAF command The result will be several files a multiple of 4 as there are 4 HWP PAs in the director
51. is provided to the principal investigator through a ftp based retrieval system Implicit in this type of observing is that no visiting observers visit the observatory rather the observations are made by the expert GTC staff astronomers The two styles of observations demand slightly different capabilities from the principal investigator This document contains all the information needed for the service observer For the classical observer the person should have read and understood this document as well as the software document For descriptions and help using such software see your local GTC staff Software such as the Java CanariCam Interface JCI and the Java Display Device JDD as well as software architecture details are described in the CanariCam Software Manual The standard classical based user should consult document CTRD 2 the CanariCam control system operator manual for JCI and JDD 4 Observing and Data Reduction 4 1 Introduction In this section we discuss the CanariCam data structure additional comments on chopping and nodding observational planning specific comments on polarimetry and then the subsequent data reduction In all the sections below a working version and knowledge of both IRAF and Starlink software suites 1s assumed and internet access to Version 1 2 2009 04 23 the Gemini mid IR resources pages is essential We note the outstanding work of the Gemini mid IR team and we make substantial use of their re
52. kets see Table 12 While a typical frame time is 20 ms the CanariCam electronics can handle frame times as short as 2 5 ms However data are not saved at this rate Rather the electronics 1s designed to coadd data from the two positions of the secondary mirror into two separate chop buffers For astronomical observations data will be typically saved at a rate of one pair every two seconds 59 Version 1 2 2009 04 23 Table 12 Common Chop Nod Terminology Term Description srcl refers to on source image in Nod position A refl refers to off source image in Nod position A difl refers to srcl refl difference of on and off source for Nod A src2 refers to on source image in Nod position B ref2 refers to off source image in Nod position B dif2 refers to src2 ref2 difference of on and off source for Nod B ii refers to difl dif2 net signal of observation from one chop nod sequence refers to accumulation of sig frames throughout observation After each chop nod sequence a new sig is added to the sig accum this is used as the primary display during observations as it shows the final result of the chop nod method sig accum 1 11 On Chip Chop amp Nod Method Another standard method exists for executing chop nod background subtraction This method 15 referred to as the on chip method and 15 demonstrated in Figure 37 with data obtained using OSCIR on the Keck II Telescope in May 1998 In the on chip
53. l extinction is significantly more transparent at mid IR wavelengths making mid IR imaging an incisive and unique probe in complex visually obscured environments In order to be able to address the broadest range of astronomical problems CanariCam obtains these images with negligible degradation to the image quality delivered by the GTC telescope Narrowband mid IR imaging will be of tremendous value in determining key properties of the dust particles and in exploiting those properties to gain further insight into the nature of the IR sources For example mid IR emission features at 7 7 8 6 and 11 3 um are attributed to polycyclic aromatic hydrocarbon molecules PAHs which are a significant constituent of the interstellar medium The absorption of UV photons by a PAH molecule results in the emission of these features the relative strengths indicating among other properties the molecular size Particularly useful 15 the fact that these features are emitted by a PAH molecule wAenever it absorbs the UV photon independent of its distance from the UV emitting source e g a star Since the PAHs are mixed in with the other dust 1ndeed the PAHs are often considered to arise from very small dust grains they should be traceable by their emission features at locations too distant from the heating source for the UV flux to heat a larger classical dust grain to a high enough temperature to emit significant mid IR continuum radiation Imaging wi
54. libration Gemini also provides a useful tool for selecting flux standards http staff gemini edu kvolk standard html This selects several stars close in time and space to the science object Target position 10 10 10 20 20 20 Nearest 6 HD82381 pos 09 31 57 58 09 42 56 8 V 5 078 12 8 652Jy Finding Nearby Standard Stars for Michelle or T ReCS Calibration Nearest 7 HD76351 pos 08 55 55 55 11 37 33 7 V 5 452 12 6 3790y Enter a target position below format as in the OT say 16 44 29 470 and 23 47 58 20 for RA and Dec Nearest 2 HD82308 pos 09 31 43 23 22 58 04 7 V 4 317 112 29 0902y Nearest 8 HD83787 pos 09 41 35 12 31 16 40 1 5 899 12 7 934 respectively Also enter the length of the observation The task will return the names of the 10 closest stellar 7 j standards one might use for photometric or low resolution spectroscopic calibration for three positions 1 the Nearest 3 HD87837 pos 10 07 54 27 09 59 51 0 V 4 381 12 17 570Jy Nearest 9 HD72094 pos 08 31 35 73 18 05 39 9 V 5 366 12 13 230Jy target position 2 same declination right ascension less by half the entered exposure time and 3 same a declination right ascension plus half the entered exposure time Nearest 4 HD94336 pos 10 53 34 85 26 12 28 0 V 7 07 12 5 323Jy Nearest 10 HD80493 pos 09 21 03 30 34 23 33 2 V 3 16 12 86 920Jy For example for a 2 hour exposure the task will look for standards at RA 1 hour less and 1 hour mo
55. lly at moderate spectral resolution at low spectra resolution fringing 1s rarely observed The fringes can be identified using fourier analysis and removed This is an interactive step in the data reduction as finding the correct frequency to reduce in the fourier transformed spectrum cannot be achieved automatically A strongly fringed spectrum taken with T ReCS is shown below 43 Version 1 2 2009 04 23 hi V 1 es V Figure 24 Fringed spectra The IRAF task msdefringe can be called to interactively fourier transform and remove the fringing from the spectra When called the task fourier transforms the spectra and plots the real and imaginary spectra as a function of frequency In the case of low resolution fringing only the highest frequency component is necessary to remove One can set the range of components that are to be removed and replace these values by either zero or an interpolated value from the end points A second option is to remove all negative values but this often fails especially 1f the baseline level if slight offset to negative Below is the fourier plot for the spectrum shown fringed above 44 Version 1 2 2009 04 23 Figure 25 Fourier transform of a fringed spectra The white curve is the real part and the red curve is the imaginary part One usually looks for the fringes in the real part VW irafterm 45 Version 1 2 2009 04 23 Figure 2
56. ly segmented mirror like GTC 1s a function of the PSF azimuth Although mid IR observations are often at or near the diffraction limit seeing can degrade and deform the image quality The image below shows a long term observation at 8 8 um of the AGN solid circles at the center of Centaurus A Radomski et al 2008 and the associated PSF object open circles The PSF is clearly variable If as 15 commonly the case the PSF standard is much bluer than the science source and a wide 28 Version 1 2 2009 04 23 bandpass is used a color correction may need to be made In the extreme case of comparing a blue 10 000 K star to that of a 50 K source the FWHM correction for observations through the broadband N filter 15 25 5 Cen A 8 8um FWHM FV HIM 5 50 00 6 20 00 6 50 00 7 20 00 7 50 00 8 20 00 8 50 00 9 20 00 UT Time Figure 13 Variation of the FWHM vs time Radomski et al 2008 4 3 CanariCam Imaging Data Reduction The standard tool for CanariCam data reduction 15 the IRAF data reduction package and the Gemini tools found at http www gemini edu sciops data and results processing software must be installed and operational In the future IRAF will evolve to pyraf and the data reduction tools will be ported to that platform The imaging data reduction consists of five main tasks or combined in one meta task listed and considered below TBACKGROUND e TPREPARE VIEW e MISTACK MIREGISTER And the
57. meta task MIREDUCE 4 3 1 TBACKGROUND This task 1s used to derive statistics on the background flux for each of the chop saveset Compromised savesets with obviously high noise can be flagged to bad and thereby excluded from inclusion in the data reduction The data can be compromised due to e Clouds e Poor seeing and or guiding e Enhanced water vapor Increased noise 29 Version 1 2 2009 04 23 The screenshot below of the EPAR of TBACKGROUND shows the input parameters with the key parameter being the sigma tolerance for bad frames Whilst this task can be useful as is common with automatic tasks it can be error prone Instead it 1s preferable to use TVIEW to investigate by hand the images where possible X xterm Image Reduction and Analysis Facility PACKAGE midir TASK tbackground inimages I Input TReCS image s outimag Output image s outpref b Prefix for output image name s rawpath Path for input raw images sigma 4 Sigma tolerance for bad frames bsetfil Bad Frame list file writeps no Write file sh_chan no Show changes to image headers logfile Logfile verbose yes Verbose status 1 Exit error status 0 good gt 0 bad scanfil Internal use onlu mode 41 fo Figure 14 EPAR of TBACKGROUND 4 3 2 TPREPARE This task is used only to reformat the data for some of the later scripts and user manipulation Since pro
58. method the chopper throw 1s set to be less than the detector array field of view so that the source remains on chip in both chopper positions 1 in both the on source and off source chopper positions as referred to in the standard method This on chip chop can be seen by examining the top row of Figure 37 where in this case the source 1s bright enough to be seen in the raw frames Note that since the source is present in both positions of the chopper the chop differenced frame will contain both a positive and negative image of the source as seen in the second row of Figure 37 The telescope beam switch is then made perpendicular to the chop throw recall in the standard technique the beam switch is parallel to the chop throw The throw for the telescope beam switch 15 also set to be less than the array field of view so that the source remains on chip Again a chop differenced image 15 formed resulting in both a positive and negative image of the source Whereas in the standard technique the beam switching resulting in a change of sign of the telescope radiative offset in this technique the radiative offset has the same sign in both chop differenced images The radiative offset 1s removed and the net signal formed by taking the difference rather than the sum of the chop differenced images In this method the net signal contains four images of the source in a box pattern with the sources located at the vertices as demonstrated b
59. nd traced using the standard IRAF package nsextract which in turns uses tasks from the apextract package by calling the f extract option The spectrum is thus extracted into a wavelength calibrated raw spectrum 2368 Version 1 2 2009 04 23 Should fringing be a problem one can de fringe the extracted spectra using fl defringe Finally a telluric correction can be applied using the fl telluric command and final wavelength and flux calibrated spectrum be output A flat field can be taken should it be needed for instance if the object is observed on a substantially different position on the array from the calibration source using a flat field screen in the sector wheel of CanariCam or of the diffuse sky or dome Whilst typically not used a bias frame can be obtained by placing all the CanariCam wheels into their blank position where possible In both cases the observations are taken in stare mode and hence the files only have one extension that contains either the flat or bias image We note that the raw flat frame 1s a good approximation to that of a blackbody spectrum at the ambient temperature of the dome instrument when the observation was taken If the fl flat flag is set to true the names of the raw data files for the spectrum to be flat fielded and bias frame subtracted must be defined as flat and bias The msflatcor takes then takes these two files as input and creates a normalized flat which in then applied to
60. ndow estimated top measured bottom 14 Version 1 2 2009 04 23 Zinc Selenide 8 12 Theoretical ZnSe 100 80 S 60 m A 40 o S 20 0 5 6 7 8 9 10 11 12 13 14 15 Wavelength um ZnSe Transmission vs Wavelength 100 00 90 00 80 00 70 00 55 60 00 5 4 50 00 Transmission 5 S 40 00 30 00 20 00 10 00 0 00 0 00 2 00 4 00 6 00 8 00 10 00 12 00 14 00 16 00 18 00 Wavelength Figure 9 Transmission curve for ZnSe window estimated top measured bottom 15 Version 1 2 2009 04 23 Potassium Bromide KBr E Ht 60 L pf LE LL LL LL EIE TLIIg BE sr LLL P LLLI IL 1 OCETETErTT Ty oo ELT Tt tt 4 4 LETT 114111 LL 441 44 00411 11 ____ 02 04 06 08 10 2 0 40 60 10 20 30 40 60 100 Transmission Wavelength um KBr Transmission vs Wavelength 82 4 80 78 76 e Transmission 74 Transmission 72 7O 68 1 1 5 10 15 20 Wavelength um Figure 10 Transmission curve for the KBr window estimated top measured bottom 2 2 Sector Wheel Outsider the dewar and just in front of dewar window is the sector wheel not shown in Figure 7 The sector wheel contains four positions one of which is open
61. nsortium VISIR VLT Mid Infrared Spectrometer Imager see http www eso org instruments visir S Similar filter available in Gemini s mid infrared imagers spectrometers T ReCS and Michelle see http www gemini edu sciops instruments michelle MichIndex html 2 7 Slit Wheel Eleven positions one of which is clear open and another blank are available in the slit wheel The nine remaining positions are populated with slit masks for use in the 10 and 20 um spectroscopic modes see Table 7 All slits are 16 mm in length spanning the entire short dimension of the detector The slit widths were selected to cover a range of seeing conditions from diffraction limited at the shortest wavelength of the N band 20 Version 1 2 2009 04 23 window to 1 04 For additional blocking the slit dekker in the aperture wheel can be used as a very oversized slit but in laboratory tests this was not needed The slit parameters are listed in Table 7 Table 7 Slit Wheel Positions 10um Slit Width Diffraction arcsec Limited Central 2 8 Wollaston Prism Slide The Wollaston prism maybe inserted into the optical beam to permit polarimetric observations in conjunction with the HWP or to insert the flip mirror see grating turret section below In all other modes the slide 1s moved to the open position thereby giving this mechanism three positions of operation More details about the Wollaston and issues concerning CanariCam polarim
62. ntensity range ztrans linear Greylevel transformation linearlloglnoneluser logfile Logfile name verbose yes Verbose status 0 Exit status O good gt 0 bad scanfil Internal use onlu mode 41 for Figure 15 EPAR of TVIEW 4 3 4 MISTACK This task collapses the full data set minus any bad flagged data into a single 320 240 MEF where extension 0 is the header and 1 1s the image The user can select if the data is co added or co averaged This is the most commonly used manner to reduce MIR data e Xi xterm Image Reduction and Analysis Facility PACKAGE midir TASK mistack inimages 52004020150118 fits Input T ReCS or Michelle image s out imag Output image s outpref s Prefix for out image s rawpath Path for in raw images frametu dif Tupe of frame to combine src ref dif combine average Combining images by averagelsum fl vari no Output variance frame logfile Logfile verbose yes Verbose status 0 Exit status O good gt 0 bad scanfil Internal use onlu mode ql ZEE for HELP 31 Version 1 2 2009 04 23 Figure 16 EPAR of MISTACK 4 3 5 MIREGISTER This task collapses the full data set minus any bad flagged data into a single 320 240 MEF where extension 0 is the header and is the image The user can select if the data is co added or co averaged The difference between MIREGISTER and MISTACK is the invo
63. odology is still in an experimental phase We note that in the case of T ReCS on Gemini South flat fields are rarely used due both to the difficulty in their production and that they may actually introduce more noise than they remove Hence at this time we do not recommend that flat fields are taken for CanariCam data 1 8 Detector Temperature Control The power dissipation of an IBC BIB array depends strongly on the background photon flux illuminating the array For example the power dissipation of the Boeing 128x128 51 5 BIB high flux device HF 16 used in the University of Florida mid IR imager spectrometer OSCIR increases by a factor of two when going from zero flux 1 blanked off to high flux illumination e g broadband N imaging Without temperature control this increase in power dissipation results in the array temperature increasing by 0 6 K and a change in detector responsivity of 30 This variation of detector response makes the repeatability of calibration measurements and f lat fielding potentially problematic if the detector is not maintained at a stable temperature The required detector temperature stability 1s on the order of a few 0 01 K In the laboratory the temperature at which the detector sensitivity 1s maximized was determined to be 9 K This will be the target value for the detector temperature control During normal operations filter changes and telescope slews can produce large thermal flux variations on th
64. ometry 1713 5 14 32 27 8 12 05 9 0 12 B81Iab 0 0 1791 5 26 17 51 28 36 26 8 1 68 B7III 1 7 2618 6 58 37 55 28 58 19 5 1 51 B21ab 2 0 3982 10 08 22 31 11 58 01 9 1 35 B7V 1 6 5056 13 25 11 58 11 09 40 8 1 04 BlIII IVD 1 6 6879 18 24 10 32 34 23 04 6 1 80 B9 5III 1 6 2421 6 37 42 70 16 23 57 3 1 90 AO0IV 1 8 2491 6 45 08 92 16 42 58 0 1 47 AlV N 1 42 Q 1 36 2891 7 34 35 9 31 53 18 1 58 AlVD 1 6 4534 11 49 03 58 14 34 19 4 2 14 A3V N 1 84 Q 1 83 4905 12 51 50 08 56 13 51 2 1 76 AOp 1 7 5793 15 32 34 14 26 52 54 7 2 21 AOV N 2 19 0 2 04 6556 17 34 56 07 12 33 36 1 2 10 ASIII 1 5 7001 18 36 56 34 38 47 01 3 0 00 AOV N 0 00 0 0 00 7557 19 50 47 00 08 52 06 0 0 77 ATV 0 2 7924 20 41 25 91 45 16 49 2 1 25 A21ae 0 8 8728 22 57 39 05 29 37 20 1 1 16 A3V 1 1 1017 3 24 19 37 49 51 40 2 1 82 F5Iab 0 4 1543 4 49 50 41 06 57 40 6 3 19 F6V 2 0 1865 5 32 43 82 17 49 20 2 2 60 FOIb 1 6 2693 7 23 48 26 23 35 5 1 84 8 0 2 2943 7 39 18 12 05 13 30 0 0 34 F5IV V N 0 76 0 0 73 3775 9 32 51 43 51 40 38 3 3 20 F6IV 1 9 6553 17 37 19 13 42 59 52 2 1 86 FiII 0 7 6615 17 44 05 10 40 06 34 9 3 02 F21ae 1 8 7264 19 09 45 83 21 01 25 0 2 89 F211 III 1 7 7796 20 22 13 70 40 15 24 0 2 24 F8lIab 0 3 437 1 31 29 01 15 20 45 0 3 63 G7IIa 1 4 509 1 44 04 08 15 56 14 9 3 50 1 6 510 1 45 23 63 09 09 27 8 4 27 G8III 2 0 854 2 54 15 46 52 45 44 9 3 94 G4III A4V 1 9 915 3 04 47 79 53 30 23 2 2 95 G8III A2V 1 0 1030
65. ommissioned TIMMI 2 on the ESO 3 6 m Reimann et al 1998 and of course T ReCS on the Gemini south m Telesco et al 1998 We note that the VLT s VISIR mid IR instrument currently uses a 256x256 Boeing array To date the Raytheon device remains the largest format mid IR array in astronomical use However this device displays several unwanted effects when exposed to a very bright source Below we discuss these effects as well as readout methods used to reduce them It may be helpful in considering these effects to keep in mind that each contiguous group of 20 vertical columns is read out through one of the 16 output channels 1 e columns 1 20 are readout through channel 1 columns 21 40 are readout through channel 2 etc In addition the first pixel read out in each channel is the lower left one as displayed in Figure 1 and the last is the upper right one 1 6 1 Single Point Sampling 51 Artifacts associated with the imaging of a bright source are most noticeable in the SI sampling mode These effects sometimes referred to as a level drop or drooping have been well documented by the COMICS team Okamoto Y K et al SPIE 2002 Sako S et al PASP 2003 For convenience we follow the labeling convention set out in those papers The artifacts can be classified into four types e Pattern 1 a gradation in brightness spanning and restricted to a single channel and diminishing in brightness on the downstream pixels in terms
66. on grating LR Ref Mirror Reference mirror for LowRes 10 and LowRes 20 grating LowRes 20 20 um low resolution grating HighRes 10 10 um high resolution grating HR Ref Mirror Reference mirror for HighRes 10 grating Table 10 Mirror and Grating Repeatability Grating Wheel Ox Ox Position pixel arcsec 0 288 E PAS Version 1 2 2009 04 23 Position to Position 0 016 The grating properties are summarized in Table 11 The spectral resolution R A AA applies to the central wavelength of the grating with the use of a slit of width Amax D where Amax is the maximum wavelength accessible to the grating 13 5 um for the LoRes 10 and HiRes 10 gratings and 26 um for the LoRes 20 grating and D 10 4 m 15 the telescope aperture diameter The predicted resolutions for the low resolution gratings are shown in Table 11 The resolution of the HiRes 10 grating is limited to lt 1260 by the 15 mm pupil image near the position of the gratings Because of the dispersion of the high resolution grating AX 0 01 um only about 1 um can be observed on the array at one time In order to cover different wavelength regions within the atmospheric window the angle of the high resolution grating 1s changed by rotating the grating turret Table 11 Grating Parameters Do LowRes 10 U c LR Version 1 2 2009 04 23 3 Software Operation Queue and Classical Observing Software associated with normal obser
67. or the objects as the standard star and hence ratio out and that the calibration source 1s taken at a distance in zenith distance from the object and the sky transmission emission may have changed between observing the two objects Sky variation can be especially severe and rapid in the Ozone 9 7um region of wavelength space If one uses the blackbody calibration routines the resulting spectrum is of arbitrary flux scaling and is normalized to at a specific wavelength in the case of the low resolution N grating In the case of an absolute spectrum the units of the out are determined by the outtype parameter where the default value 1s flux density in Jy In full detail the steps of msreduce are as follows Raw data can be corrected for the bias level and flat fielded using flat but this is rarely needed for CanariCam data The raw data files can be processed to provide stacked 1mages for the sky and difference frames us fl process e Sky emission absorption lines can then be used for wavelength calibration It is generally necessary to check the initial identifications offered by the IRAF package even for low resolution spectra as the contrast between the lines and continuum is strongly dependant on the observing conditions This is achieved using the fl wavelength option The wavelength calibration solution should then be applied to the entire field of view of the array using fl transform The spectrum can be defined a
68. over plotted on the total flux image by selecting the polarimetry toolbox in gaia and selecting the appropriate vector catalogue One can measure the degree of polarization PA polarized flux and several other parameters using this tool Much more details about polpack are found in the manual and the interested user 1s strongly recommended to read that document for more details 4 5 CanariCam Coronography Data Reduction The coronographic mode of CanariCam with respects to data reduction is a subset of imaging and hence all data reduction mechanisms envisioned for this mode are exactly the same as discussed in imaging above 55 Version 1 2 2009 04 23 Appendix A Mid IR Astronomy The Atmosphere The Earth s atmosphere is not completely transparent to mid IR radiation As shown in Figure 35 the atmospheric transmission through the mid IR regime has numerous strong absorption features caused by the Earth s atmosphere Ozone is responsible for many of these features including the strong absorption at 9 6 um while CO causes the mid IR transmission to drop to zero between 14 16 um Water vapor also absorbs in the mid IR and results in many of the absorption features such as those seen in the wavelength regime between 16 30 um At A gt 40 um the atmosphere is primarily opaque to radiation until the submillimeter regime Astronomers typically observe at wavelengths where the atmospheric transmission is the highest r
69. pose the detector to high thermal backgrounds from the sky and telescope Thus the array must be read out quickly typically at lt 20 ms so that broad band observations may be difficult or impossible with CQS e Another effect of COS is to increase the readout noise because the net signal is derived from four times the data However this has no significant effect on imaging Version 1 2 2009 04 23 or low res spectroscopy where noise is dominated by the background It may however reduce sensitivity in the high resolution spectroscopy modes Finally in order to correct for artifacts 3 and 4 one must have at least one channel not overlapping the astronomical source in order to provide a reference that can be subtracted This may be difficult in the case of extended bright sources and spectroscopic data of bright sources Further characterization of this phenomenon will be made during on sky commissioning of CanariCam on the GTC and one should check the CanariCam WWW page for the latest information 0 7 0 6 0 5 0 4 Crosstalk of peak 0 3 e 0 2 0 1 T 7 40 200 300 400 Frame time ms Figure 4 Percentage of level drop vs frame time Level drop percentage is calculated by taking ratio of source peak to level drop minimum 0 5 0 4 Crosstalk of peak i J 02 0 1 0 of e e e e e l 1 2 3 4 5 6 7 Spot H from center Figure 5 Pe
70. rcentage level drop vs spot number from center as measured in two different filters and thus two different flux levels Level drop percentage calculated by taking ratio of source peak to level drop minimum Version 1 2 2009 04 23 0 40 Crosstalk of 0 35 1 0 30 05 10 415 20 25 Filter width pm Figure 6 Percentage of level drop vs flux level as changed in the laboratory by increasing the filter bandwidth Level drop percentage 15 calculated based on taking the ratio of the source peak to that of the level drop minimum To conclude the discussion on cross talk of the CanariCam array we stress that it 1s only a noticeable effect for very bright objects While it may appear prominent the amount of negative flux is quite small and therefore affects on photometry are usually insignificant For observing faint sources near a bright object however one may want to consider use of the coronographic mode which will eliminate any cross talk problems 1 7 Mid Infrared Flat Fielding Flat field frames are needed to permit correction for fixed pattern noise in the array There is no standard method for obtaining accurate flat fields in the mid IR Inaccurate flat fields can leave residual fixed pattern noise or worse actually add noise and structure to the science data Currently flat fielding techniques are still in the development and test phase for CanariCam Other discussions of techniques used for mid IR flat field
71. re than the aen es Nearest 5 HD81146 pos 09 24 39 26 26 10 56 4 V 4 473 12 9 826Jy target position as well as using the target position Three sets of potential standards are returned Nearest 6 HD83787 pos 09 41 35 12 31 16 40 1 5 899 12 7 934 Btter standards Right Ascention 10 10 10 Declination 20 20 20 Nearest 7 HD82381 pos 09 31 57 58 09 42 56 8 V 5 078 12 8 652dy Nearest 1 HD94336 pos 10 53 34 85 26 12 28 0 V 7 07 12 5 323Jy hints Nearest 8 HD79554 09 15 13 85 14 56 29 4 V 5 360 12 5 492 Hesrest 24 8098252 pos ENIM 2B ETE 391051 9955 VEDO 121 35 6800Y Retrieve Output Clear Form Nearest 9 HD94264 pos 10 53 18 71 34 12 53 5 V 3 83 12 12 630Jy Nearest 3 HD94264 pos 10 53 18 71 34 12 53 5 V 3 83 12 12 630Jy oe Ne ea oe a Nearest 10 HD83425 pos 09 38 27 29 04 38 57 5 V 4 681 12 10 880Jy Nearest A7 BDHUOST pos o diras STOP OWOSISEERESA TOT d Ody This page was last modified on December 28 2006 Nearest 5 HD85503 pos 09 52 45 82 26 00 25 0 V 3 88 12 14 630Jy Kevin Volk kvolk gemini edu Borers Nearest 6 HD98118 pos 11 17 17 40 02 00 38 0 V 5 183 12 12 790Jy Nearest 1 HD82308 pos 09 31 43 23 22 58 04 7 V 4 317 12 29 090Jy Nearest 7 HD82308 pos 09 31 43 23 22 58 04 7 V 4 317 12 29 090Jy Nearest 2 HD81146 pos 09 24 39 26 26 10 56 4 V 4 473 12 9 826Jy Nearest
72. rom the currently selected image and the scaling to use for pixel values the map tet 16 bors 12040 1300 Rwiwere mage Feder bod coord nates c vappras Stell of acis SW ed oig any Dn Poluinelry mock oae This area displays current option settings other useful items information This area displays the currently selected image together with markers for any identified features or masks Figure 34 The POLKA graphical user interface The polka GUI 15 the crucial part of the CanariCam polarimetry data reduction Polka knows the HWP PA as it interrogates the fitsheader and with a few more inputs through the GUI the polarimetry can be trivially reduced The polka GUI 15 split into three areas the dialogue boxes the graysacle image of the polarimetric image and the status panel The suggested inputs are a In the options menu turn off sky subtraction this has already been performed in the MIR data reduction steps before Also in the options tab dual beam mode should be selected Next an o ray feature should be defined Ideally this 1s an object that can be easily centroided in any of the polarimetric slots If necessary some of the images may need to be co added to increase the source s S N in which case one should quit polka and perform this step and re start polka The o ra
73. rvation It would therefore seem that for a spatially compact source in which either method could be employed the on chip method would result in a larger source signal to noise ratio S N than the standard method given the same total observation time However due to the fact that the observations are background noise limited the signal to noise ratios are identical for a given observation time 61 Version 1 2 2009 04 23 For clarity consider a chop nod observation that consists of a single chop cycle and a single nod cycle Therefore the observations consist of the accumulation of exactly 4 frames of data Let the noise in each of the 4 frames be given by Ofame Let the source signal in a single frame be given by 5 In the standard chop nod technique each of the chop differenced frames consists of the difference between one frame with the source and one frame without Since the noise in resultant chop differenced frames adds in quadrature the S N on the single source appearing in a chop differenced frame is given by 42 Ogame The net signal is then formed by adding the two chop differenced frames together Noting that in the standard method the source images are located in exactly the same place in each of the two chop differenced frames the source signal doubles and the noise increases by V2 when the images are combined so that the final S N becomes 5 Next consider the on chip technique In this method each of
74. s certain and fewer lines are indentified leading to a worse confidence in the spectral fit 39 Version 1 2 2009 04 23 Figure 21 N band spectra during the line ID phase using data taken in wet conditions with a wide slit 26 88 74568 0089 _ 74670 43 82 78907 1222 _ 78750 106 18 94944 75 1 95030 192 82 117330 486 117280 237 00 128749 632 128770 Finally below we show the line ID for a Q band spectra In the Q band the atmospheric bands are generally stronger than in the N band In almost all cases all 13 spectral features shown below are present 40 Version 1 2 2009 04 23 Figure 22 Q band spectra during the line ID phase using data taken in typical conditions with a standard slit 4 4 4 Wavelength Transformation After the initial line identification the IRAF tasks then seek to identify the same lines in the spectra in other parts of the sky spectrum since the sky fills the slit array completely In most cases this can be achieved automatically but there are rare cases when this automation fails and it is necessary to report this process interactively Once the lines 41 Version 1 2 2009 04 23 have been identified across the spectral image a transformation 1s calculated from pixel position to wavelength and this is applied to the spectrum when it is extracted in the next step 4 4 5 Spectral Extraction The spectrum can be extracted using the IRAF NOAO twodspec package apall
75. s lost Rather one should ensure the data Is processed with particular attention paid to the HWP PA 4 5 1 Data Reduction Overview The essential data reduction steps for CanariCam polarimetry are Perform the chop nod correction as for standard imaging data reduction 2 Extract the o and e rays nothing the HWP PA used 3 Align the images This is a crucial step as a misalignment especially where there is a strong flux gradient can lead to very high spurious polarizations at highly variable PA The polarization is calculated by estimating the Stokes parameters for each resolution element The Stokes parameters used in linear polarimetry are I total intensity Q horizontally polarized and U vertically polarized a The Stokes parameters are typically estimated using standard algorithms as discussed here http www starlink rl ac uk star docs sun223 htx sun223 html and http www starlink rl ac uk star docs sun223 htx node61 htmlZAPP POL Once the Stokes parameters have been estimated the degree of polarization can be estimated using the relationships also described here http www starlink rl ac uk star docs sun223 htx node15 html 52 _ Version 1 2 2009 04 23 Ip VQ U LI 0 Q 5 arctan U Q 6 Polarimetric data is perhaps best displayed as a series of polarization vectors overlaid on a total intensity image of the source The length of the polarization vector is proportional to the degree of polarizat
76. sher Jim DeBuizer Kathleen Labrie and many others Version 1 2 2009 04 23 Science Drivers CanariCam will provide remarkable insight into the nature of a broad range of astronomical objects and environments These advances will result from the powerful combination of the world s largest optical IR telescope the 10 4 meter state of the art GTC and the high throughput excellent optics and robust mechanical structure of CanariCam that will empower the astronomer with its versatile mid IR coverage The design of CanariCam was driven by the broad range of scientific issues that it can address simply and elegantly Objects at temperatures of 100 1000 K emit significant mid IR radiation Of particular importance are the ubiquitous small solid particles dust that absorb radiation at virtually any wavelength and re radiate it as IR sub millimeter or millimeter radiation Mid IR continuum emission from the dust is diagnostic of the properties of a large variety of astrophysical objects as diverse as star forming regions circumstellar disks and starburst and active galaxies With multi wavelength mid IR imaging it is possible to locate the energy sources that power their often enormous luminosities trace the distributions of the dust particles and their temperatures and determine how UV and optical radiation which heats the dust propagates throughout the R emitting regions Furthermore the dust responsible for the often heavy visua
77. sion pixels 40 7 30 25 20 Dispersion pixels 5 10 15 20 25 Figure 31 Measured polarimetric dispersion measured in pixels versus wavelength um for the CanariCam Wollaston prism _50 Version 1 2 2009 04 23 In addition to the varying separation between the o and e rays the chromatic effect means that in broadband filters such as the N band filter there is a slight elongation of the derived PSF along the polarimetric dispersion direction It 1s expected that image quality degradation due to atmospheric conditions guiding errors or other issues will dominate over this elongation but one should still be aware of this issues as shown in figure 32 below In this example a monochromatic source at 10 5um Strehl ratio 98 is compared to a polychromatic source at 7 5 10 5 and 13 5um Strehl ratio 81 and hence is very much a worst case scenario very unlikely to be replicated in observational conditions Figure 32 Monochromatic 10 5um o ray image left and polychromatic o ray image right A further complication in polarimetry mode arises as the HWP retards the beam by half a wave at one wavelength only 10 0um That 15 the retardation is a function of wavelength with the optimal efficiency of retardation 100 occurring at a single wavelength The roll off in retardance and efficiency as a function of wavelength is shown below in Figure 33 This response was measured in the laboratory and will be re
78. sion 1 2 2009 04 23 2 If using a Si filter with a shorter than the approximate A of the N band filter the diffraction limit is smaller providing higher spatial resolution images Further considerations of filter selection include the sensitivity of the filter and the source color through the mid IR windows which bears on system sensitivity e g flux standards are often blue and can be very difficult to observe at long wavelengths flux calibration a strong color through the filter should be corrected for and PSF corrections a blue object will have a smaller observed FWHM than a red object of the same true width when observed through a filter with a finite filter width The seeing at mid IR wavelengths is much better than at optical and near IR wavelengths with mid IR imaging being typically at or near the diffraction limit 0 25 at N 0 5 at Q With good seeing and good telescope performance 2 3 Airy rings can be observed around a bright object Typical recorded Strehl ratios at Gemini are 0 6 at N and 0 9 at Q When imaging two calibrators are typically required one to monitor the point spread function PSF and one to establish the flux Typically these are two different sources because the flux standard which is bright is usually located far from the science source on the sky whereas the PSF star must be relatively close to the program source It is important to observe the PSF standard as close in time and tel
79. sults The key people associated with those activities are Drs Rachel Mason Kevin Volk and Scott Fisher We also use sections of the Starlink POLPACK software written and documented by Berry amp Gledhill Much of the IRAF script work for the software was written by or under the direction of Kathleen Labrie also at Gemini We assume that the user has 1 A Unix based computer 2 A working knowledge of IRAF essential with a working knowledge of Starlink also being desirable 3 The Gemini version of IRAF version 1 9 1 tested and running The package can be downloaded from http www gemini edu sciops data and results processing software 4 Starlink software version Humu Altair tested and running The package can be downloaded from http starlink ach hawan edu 42 Chopping amp Nodding It is generally believed that we chop 1 oscillate at several hertz the telescope secondary mirror wholly to account for the time variable sky background and or transmission While the sky variation can be significant one must make chopped observations even for perfect sky conditions due to the so called 1 f one over f noise that is common to mid IR arrays The 1 f noise arises from electronic noise within the array and associated circuitry and hence is intrinsic Each array has a characteristic frequency distribution of noise Furthermore the closed cycle cooler which keeps the array cool enough for science operation produc
80. th narrowband filters centered on PAH and other features e g the silicate features the Nel 12 8 um emission line etc presents outstanding opportunities for exploration with CanariCam The CanariCam spectroscopic mode will provide access to numerous emission and absorption features that will significantly enhance CanariCam s diagnostic power CanariCam incorporates for the first time a dual beam MIR polarimeter for use at all wavelengths in the 10 um window This will provide a uniquely powerful diagnostic of Ed Version 1 2 2009 04 23 objects as diverse as young stars debris disk and AGN Since all the polarimetry optics cool and highly transmissive 8090 polarimetry using CanariCam can be quickly implemented In this manual special attention is given to the data reduction of CanariCam polarimetry data in order to guide the non polarimetry specialist Finally a coronographic mode 15 offered for all wavelengths in the 10 um window This mode will facilitate the investigation of faint objects near to relatively bright sources such as brown dwarf candidates close to a bright parent star Through broadband and narrowband imaging long slit spectroscopy polarimetry and corongraphy CanariCam will support the exploration of a uniquely informative spectral region at a telescope that 1s well placed for this exploration Of the broad range of scientific programs that will be carried out at the GTC those in the mid IR will
81. the science data by 1 subtracting off the bias level of the raw 1mages and 2 dividing the resulting image by the normalized flat image At the conclusion of this stage an is prefixed to the file name In the case of point sources the flat fielding appears to make no appreciable difference to the resultant spectra In the case of extended sources the flat field correction 15 potentially important since the spectra of the source and standard may be observed at different parts of the array In all cases flat fielding of images 15 applied to the raw data files before any stacking and differencing of the images is executed 4 4 2 Initial Processing This process takes the raw or flat fielded data and carries out the initial processing required for all MIR data processing The task 1s called using tprepare or automatically called using the mistack task For spectroscopy one must stack the source frames in addition to make a sky spectrum images which is used for the wavelength calibration Three output files are resultant from this step 1 The prepared raw data file prefixed with a t 2 The stacked difference file prefixed with r 3 The stacked sky spectrum file prefixed with an 4 4 3 Line IDs This process identifies emission and or absorption lines in the spectra for wavelength calibration It typically needs to be accomplished in an interactive manner as in most cases the lines marked automatically
82. the sky scintillate This effect limits the angular resolution of all but the smallest optical telescopes Ideally the 1mage of a point like source in the focal plane of a telescope should resemble a classical disk In practice however time averaged images resemble a two dimensional Gaussian distribution with a full width at half max rarely smaller than 1 at optical wavelengths Seeing however is wavelength dependent with image size proportional Therefore longer wavelength mid IR observations are much less affected by this phenomenon This results in mid IR images being primarily diffraction limited Background Subtraction The primary characteristic of ground based mid infrared 5 30 um astronomy is high background photon flux The source of this background flux is the combined emission from the atmosphere the telescope mirrors and the entrance window of the cryostat that houses the camera optics and detector For example using T ReCS on GTC South with the PAH2 filter Ao 8 6 um AA 0 437 um we observe a photon flux on the detector equal to 1 13 x 10 photons pixel s This will fill the wells to 60 in 50 ms assuming a quantum yield nG of 0 40 and a well depth of 3 0x10 electrons low gain mode If the data from each pixel of the 320 x 240 array are digitized at 16 bits 1 e 2 bytes the corresponding data rate is 3 0 Mbytes s The short frame integration time therefore requires fast electronics to
83. trum is plotted 1n the upper panel Scaling can usually be found that shows the proposed output spectrum fairly well Then one needs to search the parameter space to get a good output spectrum It is generally best to change the shift value with the shift command Shifts are usually small so starting at zero and then checking values slightly positive and negative from there is usually helpful where the main goal 1s to choose a shift that removes the ozone feature as well as possible Wrong shifts cause the edges of the band to no longer match leaving a residual feature that tends to show up at around 9 3 9 6 um The following figures show 1 The initial plot screen 2 The plot screen once the wavelength range has been restricted 3 The plot screen when the shift has been set to 0 pixels In the final case no shift produces a nice smooth output spectrum over the region seen in the plot And the final result 1s the spectrum of an asteroid so what is seen 1s something that is very close to a 217 K blackbody shape 46 Version 1 2 2009 04 23 Figure 27 The initial plot with a 1 4 pixel shift as found by the telluric task d 47 Version 1 2 2009 04 23 Figure 28 The plot after changing the range and setting the offset as well as setting dscale to 0 0 irafterm efx Figure 29 The plot with a shift of 0 0 pixels 48 Version 1 2 2009 04 23 Figure 30 The final fl
84. uivalent of mireduce 1s called msreduce There are two key ways to use this package without or without executing telluric corrections through the use of the fl standard flag In this case a wavelength calibrated raw spectrum is produced If a flux standard 1s available the spectrum can be reduced to provide a flux and wavelength calibrated spectrum where the flux can be either to an absolute or relative scale To start msreduce the only parameter required 15 the object s raw data file name Of course if no telluric correction 1s needed no standard spectrum 1s needed to be specified Typically however a standard is used and the standard is defined by setting the fl standard to yes There are two possibilities set through flags if fl standard is true a If one sets the blackbody flat to yes then the spectral shape of the calibration object will be assumed to follow that of a blackbody source b Ifa standard with a known spectrophotometric energy distribution SED is used for the telluric correction the correction for both the atmospheric transmission and absolute flux level can be determined If absolute flux calibration 1s to be carried out the name of the flux standard should be given in the stdname field In the midir package SED data for the TIMMI2 and a large set of Cohen standards is available However despite carefully following these steps two uncertainties remain that of slit losses which are typically assumed to be the same f
85. umn as well as Qa 18 30um 40 23400 Jansky Qa 18 30um 3 59159E 13 changing with the weather Qb 24 56um 22 42900 Jansky Qb 24 56um 1 11498E 13 Si 1 7 73um 215 65199 Jansky Si 1 7 73um 1 08393E 11 The names of the primary and secondary Cohen standards need to be entered in specific forms such as alpha CMa to Si 2 8 74um 174 02200 Jansky Si 2 8 74um 6 88474E 12 match the internal list A list of the names of all the standard stars as used in this form can be found here Si 3 9 69um 142 19600 Jansky Si 3 9 69um 4 57099E 12 Si 4 10 38um 123 40200 Jansky Si 4 10 38um 3 43909E 12 Si 5 11 66um 98 76800 Jansky Si 5 11 66um 2 18788E 12 Si 6 12 5um 88 06000 Jansky Si 6 12 5um 1 73907E 12 Standard name ArIII 8 99um 164 16499 Jansky ArIII 8 99um 6 07961 12 NeII 12 81um 82 07400 Jansky NeII 12 81um 1 50259E 12 Instrument Both NeII cont 13 10um 78 91700 Jansky NeII cont 13 10um 1 38727E 12 SIV 10 52um 121 02600 Jansky SIV 10 52um 3 31277E 12 Airmass Zoro PAH 8 6um 178 42200 Jansky PAH 8 6um 7 22994E 12 E E PAH 11 3um 104 28800 Jansky PAH 11 3um 2 45855E 12 Watt meter square micron Output quantity Retrieve Output The values are given to more decimal places than the accuracy of the calibrations really affords Round off as appropriate This page was Last modified on May 16 2006 Kevin Volk kvolk gemini edu Done 22 p Sp Done 9 Ol 2 Done 0 5 e Figure 11 Gemini tools for in band flux ca
86. urned on and the pupil mask 1s rotated to compensate for the pupil image rotation that occurred during the previous data acquisition Obviously with the occulting mask in place any faint point or extended source around a bright object such as a star must be farther away than 0 84 arcsec from the star to be detected Simulations indicate that under good sky conditions the PSF suppression ratio the ratio of the residual point spread function PSF to the un occulted PSF can be as good as 1096 Since much or all of the noise associated with detecting a faint source near Version 1 2 2009 04 23 a bright source 1s due to PSF subtraction the ability to use coronagraphy to decrease the PSF wings by a factor of ten can reduce that subtraction noise so low that the observations can become background limited Once the above procedures associated with implementation of the masks 1s followed then the observations proceed as normal imaging observations 1 5 Basic Instrument Characteristics Detector Pixel Format 320 x 240 Science Operating Wavelength Range 7 5 26 um Pixel Scale 0 08 arcsec pixel Field of View 25 6 x 19 2 arcsec 10 um Low Resolution Spectroscopy R A AA 175 near 10 5 um 10 um High Resolution Spectroscopy R A AA 1300 near 10 5 um 20 um Low Resolution Spectroscopy R AA 175 near 20 5 um 20 um High Resolution Spectroscopy R A AA 890 near 20 5 um Coronagraphic Mode Any within the N band Polarimetry Mode Dual b
87. ux and wavelength calibrated spectrum of the asteroid Vesta 4 5 CanariCam Polarimetry Data Reduction The process of polarimetry data reduction 15 initially very similar to that of imaging data reduction but where the position angle PA of the half wave plate HWP is used to separate 1 e no combine the files The exact methodology of data collection for CanariCam polarimetry is to be defined during commissioning of CanariCam as the sky telescope and instrument will have a fundamental effect on the data collection and hence this data reduction after the following steps have been taken 1 The chop nod correction has been completed 2 Files are stored and accessible in any naming convention but that the fitsheaders are maintained as written by CanariCam 3 The files to be operated on have similar sky background image quality and noise features When using CanariCam in polarimetric mode we use the instrument in effectively imaging or spectroscopy but with the additional following components in order in the beam 1 Focal plane mask 2 Half wave plate 3 Wollaston prism 49 Version 1 2 2009 04 23 The HWP 15 used to modulate rotate the PA of polarization observed by the Wollastron prism rather than rotating the instrument A mechanical rotation of 0 results in an optical rotation of 20 and as one typically wants to sample the polarization at any PA the HWP must be mechanically rotated to a minimum of 0 and 22 5 but as a
88. ving using CanariCam is provided by the GTC Observatory This will include the observation request bidding for time phase 1 and the detailed manner in which the observations should be executed the observing tool phase 2 CanariCam will have two primary styles of observing classical or service analogous to queue observing We discuss the two styles below 1 Classical This is the traditional type of observing where the astronomer visits the telescope and executes the observations with help from the observatory staff Data is taken in the conditions at that moment and no observatory attempt 15 made to match the type of observations to the prevailing conditions If the weather is unsuitable for CanariCam observations it maybe possible to switch to another instrument but this is wholly dependant on GTC rules and regulations We note that MIR observations typically require demanding observing conditions and it is our experience that classical observing is an often poor way to collect MIR data Phase 1 and 2 are still needed to be completed and depending on GTC rules and regulations changes to the phase 2 targets and or observation mode may be disallowed 2 Service or queue In this observing style the phase 2 program is executed by observatory staff without further contact with the principal investigator The observations are executed only when observing conditions at that time meet with the principal investigator s requirements Data
89. y although the typical user would usually monitor the exposure times and clock times Because the exposure time is built from units of frametime and there must be an even number of chop nods as well as certain values for some other system parameters it works out that only certain discrete values for the exposure time are permissable Therefore the CanariCam observing software adjusts the exposure time slightly to the next longest value that fits the set of time constraints 4 3 CanariCam Filters and Calibration Objects The CanariCam filter set spans the 10 and 20 um atmospheric windows Spectrally broad gt gt lum medium lum and narrow 0 1 um filters are available offering a versatile imaging suite of filters The 10 um window is far cleaner high transparency with few absorption bands than the 20 um window However even the 10 um window suffers from a strong and highly variable absorption band from Ozone centered at 9 7 um In both atmospheric windows the atmospheric absorption band s increase the photon noise and can degrade the final signal to noise ratio of the images Interestingly in the case of the N band filter 1t has been shown on sky that the intermediate bandwidth filters of the silicate set Sil 6 have similar sensitivities to the broadband filter When using the Si filters one can gain two key advantages over and above the N band filter 1 The flux is from a more constrained wave band 06 Ver
90. y Starlink operates not on fits files but sdf exactly equivalent to ndf and hence one must convert the CanariCam fits files to CanariCam sdf files To do this once must start the conversion program and convert all files by typing 1 convert 2 fits2ndf A set of sdf files are now in the same directory as the fits files Next the polpack GUI graphical user interface should be started with the commands 1 polpack 2 polka _ 53 Version 1 2 POLKA 2009 04 23 This area allows reference objects to be displayed over the image which aid the identification of the objects currently being entered This area specifies which objects are currently being identified currently selected image This area contains a short description of the control or area currently under the pointer within iment Edit oF a features E nw fenira E ra mas ore Reference features E np y chp wee E m chp pal white y Nom ad 94 1 Draw Vaf l X hark aCe Lack Scalig Node Bh deis vue 15277 White dto vats Cun emit ution Feature 326 3 Pointer ccorcinajes megs ev Found Sty Horde 9 Shy colin ed n displayed image Specify the percentage of pida toke beach This area controls the section to be displayed f
91. y feature should be first selected on the current panel of the dialogue window and then a feature selected on the image b 54 Version 1 2 2009 04 23 c Next user should select the e ray feature on the dialogue window and a the same feature selected in the e ray image This process can be repeated for more than one source should this be available e Nextthe o ray area to be extracted should be selected Click first on the o ray mask in the dialogue box and then click on the image to define the polynomial extraction area f The e ray area to be extracted is automatically calculated from knowledge found in sections b c and d and can be shown by clicking on the e ray mask in reference g This information can be transferred to all images by using the image menu selecting all and then clicking on the transfer button h The I Q and U images are calculated when one clicks on the file menu and selects exit This process is automatic very reliable but can take some time 1 The final image output has three planes and they are the I and U images The vector map maybe produced through use of the polvec command of polpack which converts the I Q and U images to a vector map Should one want to bin the polarization vectors to achieve a higher S N this can be accomplished using the polbin command of polpack To display the eventual image the optimal image display tool is gaia In gaia the polarization vectors can be
92. y the single image in the third row of Figure 37 Two of the source images are positive and two are negative with the pattern such that the positive source are diagonally opposed and the negative sources are diagonally opposed In post processing the final image is constructed by bisecting the net signal image into 60 Version 1 2 2009 04 23 four pieces and then registering and stacking with the appropriate sign change to form the final image It is clear that the basic requirement to use this on chip method is that the program source be spatially compact less than 1 4 of the array field of view What is not so obvious is that while all of the source photons are collected with this technique as opposed to only half of the source photons collected in the standard method the source signal to noise ratio in the final image ends up being the same for both methods This 1s a consequence of operating in the background limited noise regime and 15 discussed further in the next section Nod Position A Nod Position B Net Signal Figure 37 Illustration of the on chip method Comparison of Signal to Noise for Methods In the standard chop nod method described in Section ILI only half of the total time spent accumulating photons is actually spent collecting source photons On the other hand in the on chip method described in Section source photons are accumulated for the entire duration of the obse

CanariCam User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents