Orsi by wuxiangyu

VIEWS: 5 PAGES: 54

									Lyα emitters in galaxy formation
            models
                Alvaro Orsi
                   Cedric Lacey
                  Carlton Baugh

     Supercomputing techniques in Astrophysics workshop
     Emission-line Galaxies
                        Ionizing photons
                          from young
                          massive stars

•Galaxies with
detectable emission    Absorption in HI
                       clouds in the ISM
lines

•Tracers of star         Line emission
                          from atomic
                             physics
formation activity

•UV →Lyα,Hα, Hβ,        Dust attenuation


[OII], [OIII],...
                          Observable
                          spectral line
                        Motivation
• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?
                        Motivation
                                                       Orsi et al (2009)

• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?

                                    Emission line galaxies at z=1
                        Motivation
                                                Orsi et al (2009)

• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?

                                     H-band selected
                                      galaxies at z=1
                        Motivation
                                                       Orsi et al (2009)
• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?

                                    Luminosity function of Hα emitters
                                                 at z ~ 1
                        Motivation
• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?   Euclid space mission: Hα emitters
                                          slitless survey, 0.5<z<2
                        Motivation
• Spatial distribution
- Infer mass of dark matter
  haloes hosting galaxies
• Galaxy formation in the
  high redshift Universe
- What ingredients do we need?
• Dark energy surveys
- How galaxies trace DM
  structure on very large scales?   HETDEX: Lyα emitters IFU survey at
                                                 3<z<5
                      Lyα emitters
                                     Narrow band Lyα search

• Hydrogen recombination
  line
• Strongest transition
• λ0 = 1216 Å
• Tracer of high redshift
  galaxies (2 < z < 7), aiming to
  z>7
• Resonant scattering + dust
     Small fraction of photons
     escape from the galaxy          Lyα spectrum at z=3
                                     (Gronwall et al. 2006)
         Modelling Lyα emitters
• We use the semi-analytic
  model GALFORM
  developed at Durham
• Simulate galaxy
  populations in
  cosmological volumes
• Star formation and
  galaxy merger history
  from first principles
         Modelling Lyα emitters
• Lyα emitters are modelled using the Baugh et al
  (2005) model:

- Kennicut IMF for quiescent galaxies
- Top-heavy IMF for starbursts
- SN + Superwind mode of feedback
- Monte Carlo merger trees
- Lyα emitters have a fixed escape fraction
- fesc(Lyα) = 0.02
 constant!
        Motivation for Baugh model
• Baugh et al (2005) model was
  not designed to predict Lyα
  emitters properties :

-   Submillimitre number counts
    and redshift distributions
-   Luminosity function of Lyman
    break galaxies
-   Galaxy evolution in the IR
    (Lacey et al 2008, 2009)

                                   Baugh et al (2005)
        Motivation for Baugh model
                                   Baugh et al (2005)
• Baugh et al (2005) model was
  not designed to predict Lyα
  emitters properties :

-   Submillimitre number counts
    and redshift distributions
-   Luminosity function of Lyman
    break galaxies
-   Galaxy evolution in the IR
    (Lacey et al 2008, 2009)
 Evolution of Lyα LF
                                             Orsi et al (2008)




Luminosity functions well reproduced in a
wide range of redshifts, even with a fixed
             escape fraction!
 Evolution of Lyα LF
                                             Orsi et al (2008)




Luminosity functions well reproduced in a
wide range of redshifts, even with a fixed
             escape fraction!
 Evolution of Lyα LF
                                             Orsi et al (2008)




Luminosity functions well reproduced in a
wide range of redshifts, even with a fixed
             escape fraction!
               Spatial distribution
• We combine                           Planting galaxies in
                                           an N-body
  GALFORM with the                         simulation:
  Millennium
  Simulation                Mass, positions and     Galform associates
                             velocities of DM        galaxies with DM
   – Box size: 500[Mpc/h]   haloes are recorded          halo mass
   – Mhalo > 1.72 x 1010
     M/h                    Central galaxies
                            goes to the centre
• Alternatively, N-body        of the halo
                                                       Satellite galaxies to
  merger trees can be                                   a random particle
  used                                                   within the halo
Lyα emitters at z = 0

• Low abundance due
to modest star
formation activity




     Dark matter


      Galaxies
Lyα emitters at z = 3.3

• Near peak of star
 formation activity
Lyα emitters at z = 5.7

• Star formation
decreases again
Lyα emitters at z = 8.5

• Star formation
decreases more
Clustering of Lyα emitters
              Two point correlation
              function fit by:

                                  
                           r
                   (r )   
                           r 
                            0

                    gal  b 2 dm
             b : effective galaxy bias
Comparing to observational data
        Mock catalogues of SXDS




      •Median w() of mock catalogues
Comparing to observational data
            Mock catalogues of SXDS




          •Median w() of mock catalogues
   •Idealized survey over much larger solid angle
    Comparing to observational data
                  Mock catalogues of SXDS




                 •Median w() of mock catalogues
          •Idealized survey over much larger solid angle
•Model agrees with observational measurements at 95% confidence
Comparing to observational data
      Mock catalogues for MUSYC, z=3




       Abundances and clustering properties are
       well reproduced in a constant escape
       fraction scenario!
But, is the escape fraction constant?




              Atek et al (2009)
      Empirical attempts to model fesc

Kobayashi et al (2008, 2009) semianalytic model


            f 0 simply proportional
           
  f esc   f 0e  x            dust (screen)
            f (1  e  x ) / x
            0                      dust (slab)
          N cold Z cold
  x
        ( N cold Z cold ) 0
       Empirical attempts to model fesc

Nagamine et al. (2008) SPH simulation



 Escape fraction scenario:
 f esc  f dust ( z )(1  f esc ) f IGM ( z )
                             ion




Duty cycle scenario
Samples diluted by a fixed fraction
       Empirical attempts to model fesc

Nagamine et al. (2008) SPH simulation



 Escape fraction scenario:
 f esc  f dust ( z )(1  f esc ) f IGM ( z )
                             ion




Duty cycle scenario
Samples diluted by a fixed fraction
  Detailed modelling of Lyα photons
• Monte Carlo                Some applications:
  Radiative Transfer         • Understand observed line profiles
                             -Verhamme et al. (2006,2008),
- Follow scattering of Lyα   • Surface brightness of Lyα emission in
  photons by HI atoms in      SPH galaxies
  the ISM                    -Laursen et al (2008,2009)
                             • Modelling observed DLAs
- How many of them           -Dijkstra et al (2006), Barnes et al (2009)
                             • Neutral gas fracion in the IGM
  escape (effect of dust)    - Cantalupo et al (2008)
- Lyα spectrum
Our goal
          The Monte Carlo code
1. Define the properties of the HI region
(geometry, density, temperature, kinematics, etc)


The frequency of the photon is characterised by
                           0
                      x
                           D
                             Vth 0
                       D 
                                c
             The Monte Carlo code
 2. Choose the location and direction and
     frequency of the photon
  ni  (sin( ) cos( ), sin( ) sin( ), cos( ))

3. The photon will travel an optical depth given by
                   
   P( )  1  e
    ( s)   x ( s)   d ( s)
                int             int
   s                       
         x (s)   d (s) N H  x  N d d
          The Monte Carlo code
 4. At the location of interaction, we calculate
    the probability of interacting with an H atom
                            nH  H ( x )
              PH ( x) 
                        nH  H ( x)  nd  d

5.1 If the photon interacts with dust, then the
dust albedo A tells us whether the photon was
absorbed or scattered. If absorbed, then the
photon is lost and we start over again.
                 The Monte Carlo code
5.2 If interacts with hydrogen then we compute
   the cross section:
                           1
     x   ( x)                H ( a, x )
                        D 
                                 y2
                   a         e
    H ( a, x ) 
                     ( x  y) 2  a 2 dy
The velocity of the atom parallel to the direction of the photon
depends on its frequency:
                                 u 2
                                    par
                   a         e
  f (u par ) 
             H (a, x) ( x  u par ) 2  a 2
           
  ua  uth  ubulk
         The Monte Carlo code
6. We perform a Lorentz transformation to the
   atom’s frame to compute the frequency and
   direction after the scattering
                
   x f  xi  (v  no  u par )

7. We repeat the process until the photon
escapes or is absorbed. The same is applied
to a large number of photons
Lyα photons escaping
from a static,
homogeneous sphere
       Lyα spectrum
Homogeneous, static slab
                                    Harrington (1973)
                                    analytical prediction


                           τ0=104
       Lyα spectrum
Homogeneous, static slab
                                    Harrington (1973)
                                    analytical prediction


                           τ0=104


                           τ0=105
       Lyα spectrum
Homogeneous, static slab
                                         Harrington (1973)
                                         analytical prediction


                           τ0=104


                           τ0=105

                                τ0=106
      Lyα spectrum
Homogeneous, static sphere


                             Dijkstra (2006)
                             analytical prediction
Escape fraction for homogeneous slab


                     Analytical solution for this case
                     (Neufeld, 1990)

                     No analytical solution for more
                     general cases
   Effect of outflow velocity
Homogeneous expanding sphere
                               v(r )  Hr
                                     vmax
                               H
                                    Rsphere
                                  Static case
   Effect of outflow velocity
Homogeneous expanding sphere
                               v(r )  Hr
                                     vmax
                               H
                                    Rsphere
                                  Vmax=20 km/s:
                                  Photons are slightly redshifted
   Effect of outflow velocity
Homogeneous expanding sphere
                               v(r )  Hr
                                     vmax
                               H
                                    Rsphere
                                  Vmax=200 km/s:
                                  Photons are completely redshifted
   Effect of outflow velocity
Homogeneous expanding sphere
                               v(r )  Hr
                                     vmax
                               H
                                    Rsphere
                                  Vmax=2000 km/s:
                                  The optical depth becomes so thin
                                  that photons escape very easily after
                                  being redshifted
               Next step
• Combine with GALFORM
• Choose a suitable geometry for the ISM
• Study dependence of escape fraction on mass,
  redshift, luminosities, metallicities, etc...
Measuring BAOs with
   Hα emitters
    Orsi et al (2009)
    Tracing large scale structure with
               Hα emitters
• Forthcoming dark energy
  space missions will
  measure BAOs using Hα
  emitters

• To what accuracy?

      Understand Hα
  emitters from a galaxy
  formation perspective
    Tracing large scale structure with
               Hα emitters Baugh model
                                Bower model

• Forthcoming dark energy
  space missions will
  measure BAOs using Hα
  emitters

• To what accuracy?

      Understand Hα
  emitters from a galaxy
  formation perspective
             Tracing large scale structure with
                        Hα emitters
     • The goal is to get an
       accurate P(k)
     • Different survey
       configurations determine
       P(k,z) and n(z):
     • We assess different
       configurations calculating
       the effective volume:

                                     2
             z1 n ( z ) P(k , z )  dV
Veff (k )   
             z0 1  n ( z ) P ( k , z )  dz
                                             dz
                                       
       Tracing large scale structure with
                  Hα emitters
   Error in the dark energy
   equation of state parameter:
                   1.5%
   wDE (%) 
                Veff [Gpc / h]3
   (Angulo et al, 2008)
• Hα emitters with f>1x10-16
  [erg s-1cm-2] can measure
  wDE with an accuracy of
  ~ 0.6% in a survey like
  EUCLID
• Other alternatives (H-band
  selected sample) can reach
  similar results if H(AB)<22
                                  redshift
                      Summary
• Galaxy formation models are able to predict the
  abundances and clustering of Lyα emitters using
  simple prescriptions for fesc
• Physical properties of Lyα emitters can be studied
  using RT models

• RT is needed for a more physical estimate of fesc in
  galaxy formation models
• Still work in progress...

								
To top