Formation and Clustering of High-redshift Galaxies 1. Observational Methods of Extragalactic Astrophysics Eric Gawiser Rutgers University MUSYC E-HDFS UBR composite.
Download ReportTranscript Formation and Clustering of High-redshift Galaxies 1. Observational Methods of Extragalactic Astrophysics Eric Gawiser Rutgers University MUSYC E-HDFS UBR composite.
Formation and Clustering of High-redshift Galaxies 1. Observational Methods of Extragalactic Astrophysics Eric Gawiser Rutgers University MUSYC E-HDFS UBR composite Observational astrophysics Observational astrophysics Wavelengths available for astronomy Must observe X-rays, UV, mid-infrared from space Choose to observe optical, near-infrared, microwave from space for better resolution, less background Current Astronomical Satellites • • • • • • Gamma-rays: SWIFT X-rays: Chandra, XMM UV: GALEX UV/Optical/NIR: HST Mid-IR/Far-IR: Spitzer Microwave: WMAP The World's Largest Optical/NIR Telescopes • 10m: South African Large Telescope (SALT), Keck (x2) • 8m: Very Large Telescope (x4), Gemini (x2), Subaru, Hobby Eberly Telescope (HET) • 6.5m: Magellan (x2), Multiple Mirror Telescope (MMT) • 4m: many, including Blanco at CTIO (Cerro Tololo Interamerican Observatory) Planned Larger Optical/NIR Telescopes • 8m: Large Synoptic Survey Telescope (LSST, 2013) • 11m: Gran Telescopio Canarias (GTC, 2009) • 12m: Large Binocular Telescope (LBT, 2008) • 20m: Giant Magellan Telescope (GMT, >2015) • 30m: Thirty Meter Telescope (TMT, >2015)Giant Segmented Mirror Telescope (GSMT) • ~30m: Extremely Large Telescope (ELT, >2015) • ~100m: Overwhelmingly Large Telescope (OWL, >2100?) South African Large Telescope (SALT, 10m) KwaZulu Natal, Rutgers have access Instruments up top are SALTICAM, Robert Stobie Spectrograph (RSS) SCS will use for spectroscopic follow-up of clusters discovered in SZ with ACT Telescopes on Mauna Kea Atacama Desert & Andes Mountains in Chile 8-meter Gemini telescope Gemini telescope at sunset 8-meter Gemini mirror! 8-meter Gemini primary mirror with reflected secondary Cleaning the primary mirror with… Gemini secondary Gemini instruments Charge Coupled Device = CCD Charge Coupled Device = CCD Why do SCS imaging at CTIO 4m? • Etendue Collecting_area FOV • CTIO4m+MOSAIC has 36'36' FOV so equals an 8m with 18'x18' FOV better than all 8m telescopes except Subaru! • Full figure of merit for dim point sources is Etendue*efficiency/PSF_area/background Telescopes at CTIO Data Equation: Imaging Start with sources (j) with flux densities f(,t) [ergs/cm2/s/Å], multiply by atmospheric transmission, convolve with PSF, add background, and multiply by telescope/detector/filter transmission: Yields F[ergs/s/Å] received at location r on focal plane. Number of photoelectrons in image i at pixel (x,y) results from dividing by energy per photon, integrating over wavelength, time, and pixel area, adding bias, dark current, fringing, cosmic rays, and bad pixels Typical Optical CCD Quantum Efficiency Optical Filters (UBVRIz) U B V R I z Filter only (dotted), with CCD QE and atmospheric transmission (solid) Data Reduction Pipeline: Imaging •Mask and replace bad pixels and cosmic rays •Subtract bias using overscan/zeros •Subtract dark current •Flat-field by dividing by domelamp or twilight "flat" image •Subtract fringes using "fringe template" •Assume F(r, ,t) constant over pixel, , t F(r, eff,ti) •Create and divide by "superskyflat" to get better estimate of Tqe(r) •Project onto tangent plane r(r) •Subtract background •Combine multiple exposures, ignoring masked pixels •Photometric calibration yields just the PSF-convolved sources Flat field image Fringe mask construction & subtraction Data Reduction Pipeline: Imaging •Mask and replace bad pixels and cosmic rays •Subtract bias using overscan/zeros •Subtract dark current •Flat-field by dividing by domelamp or twilight "flat" image •Subtract fringes using "fringe template" •Assume F(r, ,t) constant over pixel, , t F(r, eff,ti) •Create and divide by "superskyflat" to get better estimate of Tqe(r) •Project onto tangent plane r(r) •Subtract background •Combine multiple exposures, ignoring masked pixels •Photometric calibration yields just the PSF-convolved sources Astrometric calibration Data Reduction Pipeline: Imaging •Mask and replace bad pixels and cosmic rays •Subtract bias using overscan/zeros •Subtract dark current •Flat-field by dividing by domelamp or twilight "flat" image •Subtract fringes using "fringe template" •Assume F(r, ,t) constant over pixel, , t F(r, eff,ti) •Create and divide by "superskyflat" to get better estimate of Tqe(r) •Project onto tangent plane r(r) •Subtract background •Combine multiple exposures, ignoring masked pixels •Photometric calibration yields just the PSF-convolved sources Extended Hubble Deep Field-South, UBI composite MUSYC Extended Hubble Deep Field-South, UBI composite MUSYC Extended Hubble Deep Field-South, UBI composite MUSYC GOODS: CDF-S U-band CTIO 4m GOODS: CDF-S U-band CTIO 4m GOODS: CDF-S U-band CTIO 4m MUSYC U-band image Photometry Photometry uses knowledge of PSF in final image to determine individual source flux densities Crowded fields: simultaneous fit of multple sources convolved with PSF e.g. daophot Uncrowded fields: assume each pixel belongs to only one object, use circular apertures or adaptive-elliptical apertures (AUTO), e.g. Sextractor Optimal photometry for point sources is "PSF photometry" Noise given by: Variability: Photometric & Astrometric Once objects detected in final image, check photometry in individual images for variations Variable sources include variable stars, quasars, microlensing, supernovae, GRB afterglows - only certain stars are periodic For bright objects, check astrometry in individual images for variation Periodic variation caused by parallax, planets Non-periodic variation caused by proper motion Basic Spectrograph Design Gemini Multi-Object Spectrograph Gemini Multi-Object Spectrograph: Slitmasks and fibers Gemini Multi-Object Spectrograph: Gratings, Filters, and Camera GUIs for IMACS spectrograph on Magellan GUIs for IMACS spectrograph on Magellan GUIs for IMACS spectrograph on Magellan The Magellan Control Room Photo by S. Virani Raw IMACS data: mask image 2-d spectrum Data Equation: Spectroscopy Spectroscopy involves dispersion of photons from unmasked region, resulting in a convolution of position r and wavelength reaching position R on detector Yields F[ergs/s/Å] received at location r on focal plane. Number of photoelectrons in image i at pixel (x,y) results from dividing by energy per photon, integrating over time and pixel area, adding bias, dark current, fringing, cosmic rays, and bad pixels Data Reduction Pipeline: Spectroscopy •Mask and replace bad pixels and cosmic rays •Subtract bias using overscan/zeros •Subtract dark current •Flat-field by dividing by domelamp or twilight "flat" image •Subtract fringes using "fringe template" •Assume F(R,t) constant over pixel, t F(R,ti) •Using approximations r(R), (R), deproject to get F(r, ,ti) •Subtract background (e.g. a b-spline in position, wavelength) •Combine multiple exposures, ignoring masked pixels •Flux calibration yields just the unmasked PSF-convolved sources •Extract object spectrum (e.g. IRAF apall) Raw IMACS data: mask image 2-d spectrum COSMOS Pipeline: 2-d spectrum after bias subtraction and flat-fielding COSMOS Pipeline: after bias/flat after sky subtraction IMACS Spectrum Vmag~22.5 Photometric redshiftsSED fitting Normalization free, so N-1 d.o.f. T parametrized by star formation history, dust reddening Photo-z optimizes over templates T - but should marginalize i.e. SED fitting usually assumes z fixed - underestimates uncertainties Best results come from a wide range of wavelengths Conclusions Southern Cosmology Survey will use cutting-edge telescopes (SALT, CTIO 4m) for follow-up imaging and spectroscopy of clusters discovered in S-Z with ACT Photometric redshifts will be a key tool. They require careful data reduction, accurate photometry, complete sets of templates, and conservative interpretation of statistics Gawiser Problems 1 and 2 are now assigned - from the NEW printout