The First Galaxies: Evolution drivers via luminosity functions and spectroscopy through a magnifying GLASS

Transcription

1 Charlotte Mason (UCLA) Aspen, 7 Feb 2016 The First Galaxies: Evolution drivers via luminosity functions and spectroscopy through a magnifying GLASS with Tommaso Treu (UCLA), Michele Trenti (U. Melbourne), Kasper Schmidt (AIP), Adriano Fontana (OAR) and the GLASS and BoRG teams

2 Reioniation was likely associated with the formation of the first stars and galaxies Understand the evolution of the galaxy population to investigate their role in reioniation ionied HII CMB neutral HI Use spectroscopic properties of these galaxies to constrain IGM properties during EoR Dark Ages Epoch of Reioniation Loeb (2006)

3 UV Luminosity functions are one of our best tools for studying high galaxy populations and their evolution Rest frame UV light traces star forming galaxies Can be integrated to find the flux of ioniing photons available to reionie the universe log number density flattening at bright end? Bouwens+2015a faint end steepens? significant drop ~8 to ~? M UV Are there enough galaxies at >8 to reionie the universe? What will JWST see?

4 What drives evolution in the LF? halo mass function + star formation + physical conditions t

5 What is the simplest theoretical model to connect halo growth to star formation rate? Mason, Trenti & Treu, ApJ, 2015 minimal degrees of freedom self-consistency over redshift SFR(M h, ) ~ M h x gas accretion rate x ε(m h ) halo mass from cosmology assume gas follows DM ~ mass doubling rate from cosmology (Planck ΛCDM + ellipsoidal collapse, Sheth+2001 Lacey & Cole 1993) SF efficiency ~ M /M h fixed from calibration at one redshift via abundance matching very weakly evolving (Behrooi+2013) Trenti+20, Tacchella+2013

6 Our simple model is remarkably consistent with observed luminosity functions over 13 Gyr of cosmic time! he Astrophysical Journal, 813:21 (pp), 2015 November 1 (M) Mpc No evolution in feedback mechanisms needed, even in EoR Mason, Trenti, & Treu JWST: LFs to <14 This Work Finkelstein et al. (2015) Bouwens et al. (2015a) Bouwens et al. (2015b) Oesch et al. (2013) Figure 9. Predicted LFs at redshifts 2, 5,, 16 obtained by calibrating 8 Oesch et al. (2014) (see Section 2.3) our model with the Finkelstein et al. (2015a) LF at 5 (F15, dashed), compared to our reference calibration using the Bouwens et al. (2015a) LF at 5 22 (B15, solid). Shaded 20 regions show the 18 1σ confidence16 range, highlighting that within the uncertainty Mof UV the calibrations, the two approaches yield consistent results. Mason+2015b igure 7. Predicted UV LFs at low (upper) and intermediate (lower) redshift. Table 1 Best-fit Schechter Parameters for Model LFs also consistent with: luminosity density - stellar mass density luminosity-halo mass Redshift α M * log(φ * [mag 1 Mpc 3 ]) ± ± ± ± ± ± ± ± ± ±

7 Our simple model is remarkably consistent with observed luminosity functions over 13 Gyr of cosmic time! The galaxy LF bef The Astrophysical Journal, 813:21 (pp), 2015 November 1 Mason, Trenti, & Treu 2 3 feedback 4 5 (M ) Mpc 3 L(Mh,) - from clustering No evolution in mechanisms needed, even in EoR Harikane+2015 Mason+2015b Figure 7. Predicted UV LFs at low (upper) and intermediate (lower) redshift. JWST: LFs to < This Work Finkelstein et al. (2015) 6 Bouwens et al. (2015a) Bouwens et al. (2015b) 7 Oesch et al. (2013) Figure 9. Predicted LFs at redshifts 2, 5,, 16 obtained by calibrating Oesch et al. (2014) (see Section 2.3) our model with the Finkelstein et al. (2015a) LF at 5 (F15, 8 compared to our reference calibration using the Bouwens et al. dashed), 20 regions show the 18 1σ confidence 16 (2015a) LF at 5 22 (B15, solid). Shaded range, highlighting that within the uncertainty ofu the V calibrations, the two approaches yield consistent results. M TableLFs 1 Fig. 8. Predicted UV at high redshift. We show Best-fit Schechter Parameters for Modelalso LFs consistent with: the LFs using the calibration (see Section 2.3) at 5 * * 1 3 from Bouwens αet al. (2015b), with Planck 2015 cosmolredshift M log(φ [mag Mpc ]) luminosity density ogy (Planck Collaboration et0.1al. 2015). Points show the ± ± stellar mass density binned and ± upper limits (2013b, 0.07et al UV ±LFs 0.1 from Oesch ); et al (2015b); et 0.13 al. (2015b,a) Finkelstein 1.64 ± 0.11 ± 0.2 Bouwens luminosity-halo mass Shaded regions show the 1 confidence range ± ± ± ±

8 Faint galaxies are probably needed to reionie the universe ionied hydrogen fraction Q() Planck 2015 M lim = 12 M lim = 17 Ly emission Ly forest LAE clustering GRB damping wings Ly dark gaps QSO near ones QSO damping wings f esc = C = 1-6 log ξ ion ~ 25.2 (± 0.15 dex) Ouchi+2009, Robertson+2013, Schmidt+2014 electron optical depth () Planck 2015 all galaxies detectable galaxies M lim = 12 M lim = 17 Mason+2015b

9 Reioniation was the last major phase transformation of the universe and likely associated with the formation of the first stars and galaxies Understand the evolution of the galaxy population to investigate their role in reioniation ionied HII CMB neutral HI Use spectroscopic properties of these galaxies to constrain IGM properties during EoR Dark Ages Epoch of Reioniation Loeb (2006)

10 Is the sudden evolution in Lyα emission at >6 the smoking gun of Reioniation? fraction of LBGs with Lyα Increasingly neutral IGM? Need more data, at a wider range in luminosity, and independent sightlines Figure 4. Treu et al. (2013)

11 We are expanding the search for Lyα at >7 by exploiting the power of cluster lenses HST Grism Spectroscopy of massive clusters PI Treu, see Schmidt+2014,Treu orbits in Cycle 21 Including the 6 HFF and 8 CLASH clusters Grism Lens-Amplified Survey from Space glass.astro.ucla.edu Investigate galaxies and IGM at EoR [Schmidt+(incl CM) 2016] Environmental dependance on galaxy evolution [Vulcani+2015] Metallicity cycles in and out of galaxies [Jones+2015, Wang+in prep] SN searches, e.g. SN Refsdal [Kelly+2015] Cluster mass maps [Wang+2015, Hoag+in prep] Data released for 7/ clusters

12 observing strategy 5.6 Lyα redshift 13.0 Parallel fields Cluster Core Cluster Core (Ly [3.5,12.7] Throughput G800L F814W G2 F5W G141 F140W Wavelength [Å] Uninterrupted wavelength coverage 2 position angles to minimise contamination and better line identification Spectra of 00s of objects with m F140W < 24 Probes intrinsically faint objects due to cluster magnification Spectroscopic 1σ limits ~ 5x -18 erg/s/cm 2 (not accounting for lensing)

13 In 6 clusters, using >20 photometric selections for LBGs - 24/159 dropouts have Lyα (Schmidt+2016) - consistent with drop from ~6 can efficiently look for Lyα candidates at >6 Largest statistically well-defined spectroscopic sample of Lyman break galaxies at >6 & Schmidt et al. (2016) &

14 But higher spectral resolution is needed to confirm Lyα and constrain HST grism purity & completeness VLT KMOS large program (PI Fontana) 7 clusters ongoing until March hrs integration per source YJ band: μm ~70 >7 sources (~1/3 grism Lya) ~70 sources 1<<3 Keck DEIMOS and MOSFIRE (PI Bradač) 1 secure detection (Huang+2015) 3 more potential confirmations Flux density [erg/s/cm 2 /Å] Low 1 contaminant [OII] is resolvable s2a+ 955 TALK.pdf [OII] [µm] Flux density [erg/s/cm 2 /Å] =1.992

15 Conclusions UV LF and other global galaxy properties at 0 can be easily modelled by assuming halo growth is the dominant driver of galaxy growth Apart from dust, no evolution of physical conditions/feedback is needed! Lensing allows us to see intrinsically faint galaxies is providing the largest spectroscopic follow-up of LBGs at >6 24/159 Lya candidates in 6/ clusters consistent with significant drop from ~6 Extensive ground based follow-up is ongoing at VLT and Keck