Current Topics

1
Current Topics
Lyman Break Galaxies Dr Elizabeth
Stanway (E.R.Stanway_at_Bristol.ac.uk)
2
Topic Summary

• Star Forming Galaxies and the Lyman-? Line
• Lyman Break Galaxies at zlt4
• Lyman Break Galaxies at zgt4
• Lyman Break Galaxies at zgt7
• Reionisation, SFH and Luminosity Functions

3
LBGs at zgt7
z-drop candidates at z7
Bunker et al (2009), see also Bouwens Oesch
Castellano Wilkins etc, etc (About 20 papers in
Sep-Dec 2009)
4
Size Evolution to zgt7

• Galaxies at z7 continue to get smaller
• This scales as
• size ? (1z)-1.12 0.17, consistent with
  constant comoving sizes
• Most z7 candidates very compact
• (Oesch et al 2010)

5
The Rest UV spectral Slope
Y
J
H

• AGN have spectra described by a power law,
• L? ? ???????i.e L? ? ??????
• In the rest-frame ultraviolet, star forming
  galaxies also show power-law spectra
• The slope of the power law depends on the
  temperature of the emitting source
• This power law slope can be measured using
  broadband photometry

z
z7 galaxy
Magnitude gives the flux in J and H gt fJ and fH
Know the central wavelengths of J and H gt ?J
and ?H LJ/LH fJ/fH ? (?J?????????
6
Rest-UV Spectral Slope

• AGN have ?-1 at all redshifts
• Zero-age, star forming galaxies with normal
  stellar populations have ?-2
• Dust or age will make this slope redder (i.e.
  shallower)
• Within the LBG population the spectral slope is
  seen to evolve with z gt age evolution? Dust
  evolution?

Bouwens et al (2010)
7
Rest-UV slope at z 7 - 8
Bouwens et al (2010)

• At z7, candidate galaxies are very blue,
  particularly faint galaxies
• ?? lt -3 is very hard to explain with any normal
  (Population II) stellar population

8
Rest-UV slope at z 7 - 8

• Pop III stars are defined as having very low or
  zero metallicity
• With no metals, they have fewer ways to emit
  radiation (i.e. cool down)
• They can become hotter, and more massive
  (supported by radiation pressure)
• Hotter galaxies have bluer spectral slopes

Bouwens et al (2010)
? lt -3 slopes may indicate that z7 galaxies have
very low metallicity
9
Ensemble Properties of LBGs

• At z2-4, you can study individual galaxies in
  detail
• At z5-6, and more so at zgt7, this becomes much
  harder
• Studying an individual galaxy only tells you
  about its immediate environment
• By looking about the ensemble properties of
  galaxies you can study the universe as a whole gt
  observational cosmology
• By using a common selection method (LBGs), you
  are comparing like-for-like across cosmic time
• gt Insights into galaxy formation, the star
  formation histoy of the Universe and Reionisation

10
Luminosity Functions

• The number counts of galaxies changes as a
  function of luminosity
• This is described by a Schecter function
• N(L) dA ? (L/L)??e-(L/L) dA
• At low-z this parameterises the galaxy mass
  distribution
• The function has three important parameters
• Characteristic luminosity, L or M (26.5 at
  z6)
• Faint end slope, ????????
• Normalisation, ?
• (Bouwens et al, 2007)

z4
z5
z6
11
Luminosity Functions
z4
z3
z6
z5

• Number counts are affected by incompleteness and
  contamination
• There is degeneracy in the parameter fitting
• gt The exact values at high z are still
  uncertain, but
• The typical magnitude of the population is
  decreasing at high z gt younger, smaller galaxies
• The faint end slope appears steeper at high z gt
  more faint galaxies compared to bright galaxies
• At any given luminosity there are fewer z6
  galaxies than z3 galaxies
• (Bouwens et al, 2007)

12
LF Results at z gt7

• At zgt7 there are fewer galaxies and we dont
  probe the faint end slope
• The Luminosity function is continuing to evolve -
  there are fewer Lyman Break galaxies as you move
  to higher redshifts, but the fraction that are
  faint increases

Bouwens et al (2010)
13
Luminosity Function Implications

• At earlier times, star formation in the Universe
  is increasingly dominated by small, hard to
  detect galaxies
• The fraction missed by a magnitude limited survey
  is increasing
• The more massive galaxies we see at z3 are
  increasingly rare at higher z - star formation is
  occuring in less massive, less mature regions
  (i.e. lower metallicity? less dusty?)
• A Schecter function still describes the
  distribution reasonably well out to z6 - star
  formation may still be tracing the mass
  distribution despite the short-lived starbursts
• Models for hierarchical merging suggest that the
  typical luminosity is evolving to follow the
  typical galaxy mass at a given redshift

14
Cosmic Evolution of Star Formation
Property z1-3 z5-6 zgt7
Age 200 Myr 50 Myr May be younger
Mass few x 1010 M? 109 M? No data
Metallicity 0.3-0.5 Z? 0.2 Z? May be very low - Pop III
Size (half light radius) 1.5-2 kpc 1kpc scales as comoving 0.5 kpc
M -21.1 z5 -20.7 z5 -20.2 -19.9?
Faint end Slope -1.6 may be steeper No data
Dust E(B-V)0.2 Probably less dusty No data
Star Formation Rate 30 M?/yr 30 M?/yr 30 M?/yr
15
The Star Formation History of the Universe

• LBGs are star forming galaxies
• If there was other star formation at the same z
  it would be detected UNLESS it is extincted
• So LBGs can be used to measure the star formation
  history of the universe modulo dust extinction

16
The Star Formation History of the Universe

• This was first done using LBGs in the mid-1990s
  using Lyman Break Galaxies at z3-4 by Piero
  Madau
• As a result, the Star Formation History of the
  Universe is usually shown on a diagram known as
  the Madau Plot
• Early work showed that star formation peaked
  around z1, but it was unclear what happened at
  higher redshifts

(Steidel et al 1999)
17
The SFH out to z6
GOODS extended this work to z6 (for bright
galaxies)
The Star Formation Rate Density out to z6 shows
steep evolution, particularly when only bright
galaxies are considered
18
Uncertainties in the Madau Plot

• To get to star formation rate density you need
• Number of objects per unit volume
• Star formation rate per object
• You have
• Number of galaxies (Complete sample?
  Contamination?)
• Rest-UV flux (after dust extinction)
• Redshift selection function (Survey and model
  dependent)
• Uncertainties
• How much UV flux has been absorbed by dust?
• How much is emitted by galaxies below your
  selection limit?
• How do star formation rate and rest-UV flux
  relate?

19
The Star Formation History of the Universe
The LF has a steep faint end slope at high-z The
fainter you integrate down the Luminosity
Function, the more flux youll see Even to faint
magnitudes the SFRD is still dropping at high
redshift
(Bouwens et al, 2007)
20
The Star Formation History of the Universe

• Significant uncertainty in star formation density
• General consensus
• The SFRD is either steady beyond z2 or declining
  slowly
• It declines rapidly beyond z6
• Metallicity, age and duty cycle are all important
  parameters

Verma et al, 2007
21
But is this a complete picture?

• Until recently most models predicted SF peaking
  much earlier
• LBGs are selected to be rest-UV bright
• How complete is the picture of the z3 universe
  they paint?
• Do they even map out all the star formation?
• What about UV-dark or dusty material?

Many Models predict SFR should peak at zgt6
(Springel Hernquist, 2003)
22
Sub-mm and Dust-obscured Galaxies

• Rest-UV flux is reprocessed and reemitted in the
  far-infrared by dust
• At z2, 25 of the far-infrared luminosity in the
  universe is seen in IR-detected DOGs
• Luminous sub-mm galaxies are rare but can have
  SFRs of 100s of M?/yr

• Numbers of SMGs are known to peak at z2-3 (the
  epoch of galaxy mergers)
• At these redshifts, dust obscured galaxies
  might contribute 50 or more of the cosmic SFRD

23
Including Obscured Galaxies in the Madau Plot

• Newer SFH models include feedback from QSOs and
  gas
• Models tend not to consider dust obscuration
• Predict SFH peaking at z3-4
• Possible that SMGs could contribute to this
  picture, particularly at z1-3

Nagamine et al (2009)
24
Reionisation

• Lyman break galaxies are star-forming so directly
  measure the star formation properties of the
  universe
• At z7 they are starting to probe a transition
  known as reionisation when the galaxy went from
  largely neutral to largely ionised

25
Reionisation and the End of the Dark Ages

• After the Big Bang, the universe cooled and
  recombined leaving a neutral universe. At this
  time, all rest-frame UV light is absorbed by
  neutral hydrogen ? the Cosmic Dark Ages
• The IGM in the local universe is highly ionised
• So when and how did the universe reionise? What
  ended the Dark Ages?
• The WMAP satellite studied the optical depth to
  the CMB data which suggests that zreion9-12
• The failure of the LAE Luminosity Function to
  evolve implies zreiongt6
• However, the observation of a Gunn-Peterson
  trough in the spectra of SDSS z6 QSOS implies
  zreion6.2 (or at least a large neutral fraction)
• Could z6 starbursts contribute to reionisation?

26
Reionisation - Evidence from WMAP

• The CMB has been streaming through the Universe
  since Recombination
• The mean free path of CMB photons will depend on
  the distance the radiation travels through a
  neutral vs ionised medium
• WMAP has measured the CMB power spectrum,
  constraining cosmological properties

• One of these, ?, is the optical depth of CMB
  photons to reionisation
• After five years of data, the best fitting value
  suggests zreion10.8 1.4

27
Reionisation - Evidence from SDSS

• Damping of the spectrum due to Lyman-alpha forest
  lines rapidly increases with increasing redshift
• This can be seen in the spectra of distant QSOs
  seen in the SDSS
• Beyond z6.4 large regions of the spectrum are
  seen with zero flux
• These are known as Gunn-Peterson troughs and
  indicate that the universe is at least partly
  neutral beyond z6

28
Reionisation - Evidence from LAEs

• Gunn-Peterson absorption (i.e. due to neutral
  gas) has broad damping wings
• Therefore if theres neutral gas surrounding a
  Lyman-alpha emitter, the line can be suppressed,
  even though its longwards of 1216(1z) A
• In a neutral universe you expect to see a smaller
  number of Lyman-alpha emitters detected

29
Reionisation - Evidence from LAEs

• Gunn-Peterson absorption (i.e. due to neutral
  gas) has broad damping wings
• Therefore if theres neutral gas surrounding a
  Lyman-alpha emitter, the line can be suppressed,
  even though its longwards of 1216(1z) A
• In a neutral universe you expect to see a smaller
  number of Lyman-alpha emitters detected
• There is some evidence for this at z6.6

Pure num. evolution
Pure lum. evolution
30
Ouchi in prep
Lya Luminosity Function (Lya LF)
30
Reionisation - Evidence from LBGs

• The neutral IGM is ionised by UV-flux
• The dominant source of UV-flux in the universe is
  star formation
• The effect of UV-flux on the universe depends on
• Clumpy IGM
• Escape of UV-photons
• Temperature of IGM
• Cosmology
• Can determine a critical Star Formation Density
  that will ionise the Universe given some values
  for these parameters
• By integrating the LBG LF we measure the total UV
  flux and can compare it with this critical values

31
Reionisation - Evidence from LBGs

• Can calculate the ionised fraction of the
  universe due to contribution of LBG galaxies
  given certain assumptions
• For reasonable assumptions about a warm IGM, it
  is possible to fit the data (within errors) and
  ionise the universe with z7 LBGs
• BUT LBGs are scarcer at higher redshifts - is
  this a problem at z10?

Current best fit ?0.087-0.017
Oesch et al 2010
32
Lecture Summary

• With increasing redshift see
• Decreasing metallicity
• Decreasing dust extinction
• Decreasing age
• Decreasing mass
• Very blue rest-UV spectra are hinting at changes
  in the nature of star formation
• LBGs at every redshift are used to characterise
  evolution in star formation density and the
  mechanisms and environment for star formation
• This could be critical for understanding the star
  formation history of the Universe and Reionisation