Emission Line SurveysLecture 1

Mauro GiavaliscoMauro GiavaliscoSpace Telescope Science InstituteSpace Telescope Science Institute

University of Massachusetts, AmherstUniversity of Massachusetts, Amherst11

11From January 2007From January 2007

Outline DefinitionsDefinitions

Why emission linesWhy emission lines Types of surveys and methodologyTypes of surveys and methodology

Target surveysTarget surveys Blind surveysBlind surveys

SensitivitySensitivity Narrow-band imagingNarrow-band imaging Slit spectroscopySlit spectroscopy Slitless spectroscopySlitless spectroscopy

Observational techniquesObservational techniques Results from Emission Line SurveysResults from Emission Line Surveys

Historical notesHistorical notes Discussion of recent and ongoing surveysDiscussion of recent and ongoing surveys

MethodologyMethodology ResultsResults

Future prospectsFuture prospects

Disclaimer We wrote these lectures from the point of view of the We wrote these lectures from the point of view of the

“observer”“observer” They do not aim at providing a complete review of emission They do not aim at providing a complete review of emission

line surveys and their resultsline surveys and their results Rather, the choice of material is aimed at maximizing Rather, the choice of material is aimed at maximizing

pedagogical value, illustrating current interesting problems, pedagogical value, illustrating current interesting problems, and at helping potential observers planning and designing and at helping potential observers planning and designing their own emission line surveystheir own emission line surveys

It also reflects our personal tastes and biasIt also reflects our personal tastes and bias Readers are strongly encouraged to do further, comparative Readers are strongly encouraged to do further, comparative

research in any specific subject discussed here research in any specific subject discussed here

Why Emission Line Surveys To To effectivelyeffectively look for a specific class of sources in some look for a specific class of sources in some

pre-assigned volumepre-assigned volume of spaceof space and/or at some and/or at some pre-assigned pre-assigned point in timepoint in time

““effectively”: effectively”: with high yield (low contamination) and in with high yield (low contamination) and in large numberslarge numbers

Exploit the presence of emission line in the spectral Exploit the presence of emission line in the spectral energy distribution of most astrophysical sourcesenergy distribution of most astrophysical sources

Traditional flux selection plus follow-up spectroscopy Traditional flux selection plus follow-up spectroscopy highly inefficient to cull special classes of sources from highly inefficient to cull special classes of sources from the general countsthe general counts

Notations, Definitions, Reminders and World Model. I

Throughout these lectures, we use:Throughout these lectures, we use: F: flux, in units of erg/s/cmF: flux, in units of erg/s/cm22

ff: flux density, in units of erg/s/cm: flux density, in units of erg/s/cm22/Hz/Hz ff: flux density, in units of erg/s/cm: flux density, in units of erg/s/cm22//ÅÅ

ffff •• |d |d/d/d = f = f •• c/ c/22

1 1 ÅÅ = 10 = 10-8-8 cm cmc = 2.9979 c = 2.9979 • • 10101010 cm/s cm/s

Notations, Definitions, Reminders and World Model. II

Throughout these lectures, we use:Throughout these lectures, we use: L: luminosity, in units of erg/sL: luminosity, in units of erg/s ll: luminosity density, in units of erg/s/Hz: luminosity density, in units of erg/s/Hz ll: luminosity density, in units of erg/s/: luminosity density, in units of erg/s/ÅÅ

ff = l = l • (1+z)• (1+z) / 4/ 4 • D• DLL22(z)(z)

ff = l = l / 4/ 4 • D• DLL22(z) • (1+z)(z) • (1+z)

F = L / 4F = L / 4 • D• DLL22(z) (z)

• DDLL(z) = D(z) = DLL(z; (z; HH00, , mm, , ) : luminosity distance) : luminosity distance• z is the redshift defined as z = a(tz is the redshift defined as z = a(t00)/a(t) – 1)/a(t) – 1• t is the cosmic time and tt is the cosmic time and t00 is the age of the universe is the age of the universe

Notations, Definitions, Reminders and World Model. III

Throughout these lectures, we use:Throughout these lectures, we use: AB magnitudes: AB magnitudes:

mmABAB = -2.5 = -2.5 • • LogLog1010(f(f) - 48.595 ) - 48.595 • (Oke 1974; Oke & Gunn 1977)(Oke 1974; Oke & Gunn 1977)

ST magnitudesST magnitudesmmSTST = -2.5 = -2.5 • • LogLog1010(f(f) - 21.1) - 21.1

• (Walsh 1995)(Walsh 1995) World Model (when needed):World Model (when needed):

HH00 = 70 km/s/Mpc = 70 km/s/Mpcmm = 0.3; = 0.3; = 0.7 = 0.7

CCD and near-IR Detectors Most common devices used in emission line surveysMost common devices used in emission line surveys Photon counting devices:Photon counting devices:

DN = G DN = G • • NN DN: Calibrated Data Number, I.e. what we read from the detector after DN: Calibrated Data Number, I.e. what we read from the detector after

calibrationscalibrations G: inverse gainG: inverse gain NN: number of photons, in a finite wavelength interval : number of photons, in a finite wavelength interval

Detectors add their own “signal” and noise:Detectors add their own “signal” and noise: DNDNobsobs = DN + K + = DN + K +

is removed during calibration (bias + d.c. + …)is removed during calibration (bias + d.c. + …) is a random variable with is a random variable with

• <<• < < ronron22

rmsrms + d.c. + d.c.22rmsrms + … + …

• Typical values:Typical values:– [ron[ron22

rmsrms]]1/21/2 ~ a few (as low as ~1) to a few 10 e ~ a few (as low as ~1) to a few 10 e- - /pix/pix– [d.c.[d.c.22

rmsrms]]1/21/2 ~ 0.01 to a few e ~ 0.01 to a few e--/sec/pix/sec/pix Let’s assume G=1 in the followingLet’s assume G=1 in the following

The Finite Resolution element The smallest spatial scale or wavelength interval the instrumentation The smallest spatial scale or wavelength interval the instrumentation

can resolve:can resolve: Spatial (PSF): the seeing (ground) or diffraction limit (space)Spatial (PSF): the seeing (ground) or diffraction limit (space)

• Good (bad) seeing: 0.6 (2) arcsecGood (bad) seeing: 0.6 (2) arcsec• HST resolution (V band): 0.03 arcsecHST resolution (V band): 0.03 arcsec

Depends on the size of the telescope, wavelength and… luck!Depends on the size of the telescope, wavelength and… luck!• Poor image quality spreads photons over a large area, adds Poor image quality spreads photons over a large area, adds

noise (2x seeing = 4x noise)noise (2x seeing = 4x noise) Spectroscopic (resolution): the spectral resolution elementSpectroscopic (resolution): the spectral resolution element

Depends on the dispersion of the spectral element (prism, Depends on the dispersion of the spectral element (prism, grism, grating) and on the slit aperturegrism, grating) and on the slit aperture

If pixel size is well matched to resolution element (Nyquist If pixel size is well matched to resolution element (Nyquist sampling): FWHM (of PSF or LSF) covered by 4 pixelssampling): FWHM (of PSF or LSF) covered by 4 pixels

S/N: Signal-to-Noise Ratio

Most important metric to asses sensitivity.Most important metric to asses sensitivity. S/N in some finite wavelength interval S/N in some finite wavelength interval either the passband either the passband

width or the spectral resolution element width or the spectral resolution element since we detect (count) photons, uncertainty on photon counting since we detect (count) photons, uncertainty on photon counting

is simply is simply = N = N1/21/2, and thus:, and thus:

S/N = SS/N = S / [S / [S + B + B + N + N22]]1/21/2

SS: number of photons from source: number of photons from source BB: number of photons from background: number of photons from background NN: equivalent number of photons from additional sources of : equivalent number of photons from additional sources of

noise (typically detector)noise (typically detector)

Width of Emission Lines

The finite width of an emission line along the wavelength axis.The finite width of an emission line along the wavelength axis. Commonly measured by the Full Width at Half Maximum Commonly measured by the Full Width at Half Maximum

(FWHM). For a gaussian line profile:(FWHM). For a gaussian line profile: ~ 0.425 ~ 0.425 •• FWHM FWHM

The line width reflects the kinematics of the emission region The line width reflects the kinematics of the emission region (kinematics of the gas or of the individual sources in the case of (kinematics of the gas or of the individual sources in the case of integrated emission). If v is a measure of the velocity field within integrated emission). If v is a measure of the velocity field within the emission regionthe emission region / / = = v / cv / c

If source is at redshift z, wavelengths are “stretched” by (1+z), thus If source is at redshift z, wavelengths are “stretched” by (1+z), thus observed FWHM and rest-frame FWHM related by:observed FWHM and rest-frame FWHM related by: FWHM(obs) = FWHM(rest) FWHM(obs) = FWHM(rest) • • (1+z)(1+z)

Equivalent Width of Emission Lines

Metric to asses the strength of an emission line.Metric to asses the strength of an emission line. The width of a top-hat emission line of equal luminosity and peak The width of a top-hat emission line of equal luminosity and peak

value equal to the continuum at the line wavelengthvalue equal to the continuum at the line wavelength It represents the wavelength range over which the continuum It represents the wavelength range over which the continuum

luminosity equals the line luminosityluminosity equals the line luminosity

WW = L / l = L / l = F / f = F / f

Unaffected by extinction (line and continuum extinct by equal amount)Unaffected by extinction (line and continuum extinct by equal amount) If source is at redshift z, wavelengths are “stretched” by (1+z), but If source is at redshift z, wavelengths are “stretched” by (1+z), but

luminosity (number of photons) is conserved. Thus, observed Wluminosity (number of photons) is conserved. Thus, observed W and and rest-frame Wrest-frame W related by related by WW(obs) = W(obs) = W(rest) (rest) • • (1+z)(1+z)

How to Detect Emission Lines

Directly: observing the spectra of some Directly: observing the spectra of some class of candidatesclass of candidates

Indirectly: comparing the photometry of the Indirectly: comparing the photometry of the line through narrow-band passbands (on-line through narrow-band passbands (on-band images) to that of the continuum band images) to that of the continuum through either narrow or broad-band through either narrow or broad-band passbands (off-band images) passbands (off-band images)

How to Detect Emission Lines: Spectroscopy

Spectroscopy

Lyz=5.65

Vanzella et al., in prep.

How to Detect Emission Lines: Photometry

Finding galaxies at high-redshift: color selection

B435 V606 z850

Unattenuated Spectrum Spectrum

Attenuated by IGM

B435 V606 i775 z850

z~4

1. Color selection is very efficient in finding galaxies with specific spectral types in a pre-assigned redshift range

2. Wide variety of methods available, targeting a range of redshifts, galaxies’ SEDs:

• Lyman and Balmer break (Steidel et al., GOODS)

• BX/BM (Adelberger et al.,

COSMOS)• DRG (van Dokkum et al.,

GOODS)• BzK (Daddi et al.)• Photo-z (Mobasher et al)

Here, the case of “Lyman-break galaxies”

How to Detect Emission Lines: Photometry

How to Detect Emission Lines: Photometry

Source Selection

Source Selection

Emission-Line Sources

Weeding Out Interlopers

Spectroscopy Follow-up: the Interlopers

Spectroscopy Follow-up: the Targets

Spectroscopy Follow-up: the Targets