Physical Properties of Spectroscopically Confirmed Galaxies at Redshift ≥ 6

Linhua Jiang(Arizona State University )

Hubble Fellow Symposium 2013

Collaborators: E. Egami, X. Fan, S. Cohen, R. Dave, K. Finlator, N. Kashikawa, M. Mechtley, K. Shimasaku, and R. Windhorst

Outline Introduction

UV and Lyα properties• Rest-frame UV continuum slope• Rest-frame Lyα equivalent width (EW)• UV luminosity function of Lyα emitters (LAEs)• SED modeling• LAE vs. LBGs• ……

Morphology• Rest-frame UV continuum morphology• Merging/interacting systems• Lyα morphology of LAEs• Positional differences between UV continuum and Lyα emission• Nonparametric measurements• …… (Jiang et al. 2011)

z ~ 0 galaxy,1 min int. with SDSS

4’

10”

29 hr int. with Subaru (z band)

4 hr int. with Keck/DEIMOS

Epoch of Cosmic Reionization Cosmic Reionization

• Neutral IGM ionized by the first luminous objects at 6 < z < 15• Evidence: CMB polarization (Komatsu 2011) + GP troughs (Fan 2006) + …• Responsible sources for Reionization? (e.g. Finlator 2011)

High-z galaxies (z ≥ 6)• HST + the largest ground-based telescopes• A few hundred galaxies or candidates at z ≥ 6; many at z ≥ 8 (e.g. Oesch 2010; Bouwens 2011; Stark 2011; Yan 2011; and more papers in 2012 and 2013)

(Robertson+2011)

(Robertson 2011)

Note: See Dan Stark’s talk

for EoR

Galaxies at z ≥ 6 Current studies of galaxies at z ≥ 6

• Mostly done with photo-selected Lyman-break galaxies (LBGs)• Lack of a large spec-confirmed sample with deep infrared data• LBGs found by HST are faint often no spec-z, no Spitzer detections• LBGs or LAEs found by ground-based telescopes are bright, but

do not have deep near-IR (HST) or mid-IR (Spitzer) data

Our sample of spec-confirmed galaxies• A total of 67 galaxies, from the Subaru Deep Field (SDF) • 22 LAEs at z ≈ 5.7; 16 LBGs at z ≈ 6; 28 LAEs at z ≈ 6.5; 1 LAE at z ≈ 7• 3 HST and 2 Spitzer programs (PIs: L. Jiang and E. Egami)• First systematic study of spec-confirmed galaxies at z ≥ 6

Note:• LAEs: found by the narrow-band (NB)

technique

• LBGs: found by the dropout technique

See Chun Ly’s talk

for the SDF

Imaging data• Optical data from Subaru/Suprime-Cam (PSF≈0.6”)• Broad-band data: BVRi ≈ 28.5, z ≈ 27.5, y ≈ 26.5 (AB mag at 3σ)• Narrow-band data: NB816, NB921, NB973• Near-IR data: HST/WFC3 F125W (or F110W) and F160W (2 orbits per band)• Mid-IR data: Spitzer IRAC 1 and 2 (3 – 7 hrs)

z J H IRAC1 IRAC2

Rest-frame UV and Lyα properties

UV continuum slope β (fλ ~ λβ)

–1.5 ≤ β ≤ –3.5; median β ≈ –2.3 Steeper than LBGs in previous studies (β ≈ –2 ~ –2.1 ) No obvious β – M1500 relation LAEs do not have steeper β than LBGs

Extremely blue galaxies (β ~ –3!) Statistically significant excess of galaxies with β ~ –3 Nearly zero dust and metallicity + extremely young + … Current simulations cannot produce β = –3 (Finlator 2011) How to: top-heavy IMF, high escape fraction, etc. (e.g. Bouwens 2010) To confirm β ~ – 3: new HST observations

Rest-frame Lyα EW

• EW: 10 ~ 300 Å• EW – M relation and selection effect

MUV

EW

UV luminosity function (LF) of LAEs with EW > 20Å

• In the brightest end: comparable to the LF of LBGs a large fraction of LAEs among bright LBGs• In the faint end: much lower faint galaxies with weak Lyα emission could be the dominate contribution to the total ionizing flux at z ~ 6

z ≥ 7 (Oesch 2010)

4’ 4” 2” 1”

z<0.1

z=1

z=3z=6

Morphology and structure Galaxies at z ≥ 6 are faint and small

• At higher redshift: size evolution, (1+z)4 surface brightness dimming• At higher redshift: less well developed, more peculiar• Galaxies at z ≥ 6: occupy several pixels even in HST images

Galaxy size

Half-light radius rhl

• rhl ≈ 0.1”–0.28” (0.6–1.6 kpc) • Median rhl = 0.16” (1 kpc)• Intrinsic rhl,in ≈ 0.05”–0.26” (0.3–1.5 kpc)• Median rhl,in = 0.13” (0.8 kpc)• Small and compact

Size-luminosity relation• Brighter galaxies are larger • rhl ~ L0.14 and rhl,in ~ L0.20

Comparison with previous work• Similar to photo-selected LBGs• In particular, LAEs are not smaller

Nonparametric methods

Morphological parameters• CAS (Conselice 2003)• Gini and M20 (Lotz 2004,2006)

Our measurements• Different locations in parameter space,due to small galaxy sizes

Applicability• Correlation between parameters• Still applicable, if used with caution?

z=6.96 LAE

Interacting/merging galaxies

Merging galaxies• Visual identification at M1500 < –20.5 mag• ≥ 2 distinct cores, or, extended with long tails

Merger rate• 36% – 48% at M1500 < –20.5 mag• 39% – 56% at M1500 < –21 mag

Lyα morphology

Diffuse Lyα halos At low redshift: controversial At high redshift: predicted by simulations Based on ground-based NB (Lyα) images

2 < z < 3(Steidel 2011)

z = 3.1 LAEs(Matsuda 2012)

z = 3.1 LAEs(Feldmeier 2013)

z = 5.7 LAEs(Zheng 2011)

20” x 20”

No Lyα halos found in z=5.7 and 6.5 LAEs Stack 43 LAEs at z=5.7 and 40 LAEs at z=6.5 Stacked images: resolved but not very extended Possible reasons: dust, halo distribution, or no halos, etc.

Lyα–continuum misalignment

Lyα and continuum emission• Originated from the same SF regions• Lyα affected by resonant scattering

Data• UV continuum: HST • Lyα: narrow-band images

Results• Compact LAEs: no offsets• Interacting systems: a variety of positional differences• Disturbed ISM distribution

Summary

First systematic study of spec-confirmed galaxies at z ≥ 6 Steep UV continuum slopes UV continuum LF of LAEs Galaxy size: mostly small and compact Large fraction of merging galaxies in the bright end No diffuse Lyα halos were found around LAEs Significant positional differences between Lyα and UV continuum For more information:

• Jiang et al. 2013a, ApJ submitted (arXiv: 1303.0024)• Jiang et al. 2013b, ApJ submitted (arXiv: 1303.0027)

Voted as No. 1 and 2 by on Monday