Galaxy Build-Up and Evolution in the First 2 Billion Years of the Universe
Total Page:16
File Type:pdf, Size:1020Kb
Galaxy Build-up and Evolution in the First 2 Billion Years of the Universe Rychard Bouwens Leiden Observatory Special thanks to my collaborators Pascal Oesch, Ivo Labbe, Garth Illingworth, Renske Smit, Benne Holwerda... Santa Cruz Galaxy Formation Workshop 2014 Santa Cruz, California August 11, 2014 State-of-the-Art Infrared Instrumentation is allowing for Great Progress WFC3 camera on the IRAC Camera Very Wide-Area Hubble Space Telescope Spitzer Space Ground-based Telescope Cameras WFC3/IR HST Spitzer HAWK-I VISTA -- 4 arcmin field of view (6x larger than NICMOS) -- excellent sensitivity (3-4x better than NICMOS) -- excellent spatial resolution (2x higher than NICMOS) State-of-the-Art Infrared Instrumentation is allowing for Great Progress WFC3 camera on the IRAC Camera Very Wide-Area Hubble Space Telescope Spitzer Space Ground-based Telescope Cameras WFC3/IR HST Spitzer HAWK-I WFC3/IR increased search efficiency by 40x VISTA -- 4 arcmin field of view (6x larger than NICMOS) -- excellent sensitivity (3-4x better than NICMOS) -- excellent spatial resolution (2x higher than NICMOS) HUDF NICMOS J110+H160 144 orbits 4 HUDF WFC3/IR Y105+J125+JH140+H160 4 z > 6.5 galaxies (before WFC3/IR) (first 850 Myr of universe) 120 z > 6.5 galaxies (after WFC3/IR) (first 850 Myr of universe) ALL FIELDS 15 z > 6.5 galaxies (before WFC3/IR) 255 orbits > 800 z > 6.5 galaxies (after 5WFC3/IR) Why are studies of galaxies at very high redshifts interesting? -- It is when galaxies first form... (halos of L* and sub-L* galaxies built up from z~30+ to z~3) -- It is when the universe was reionized... (galaxies are most likely driver, so by studying the formation of first galaxies perhaps we can gain insight) Focus of this Presentation: Galaxy Growth from z~10 to z~4 What are the different regimes to study galaxy growth? “Normal” Population of Faint Galaxies (Most stars in universe form here) How rapidly do faint (low mass) galaxies grow up? ⇒ Study using the Hubble Ultra Deep Field (and similar fields) Rare Population of Bright Galaxies How rapidly do bright (massive) galaxies grow up? How Bright / Massive can Galaxies Become? ⇒ Study using very wide-area fields What can we learn based on current wide-area fields? Full CANDELS Program (+ BORG + ERS) provides an ideal dataCANDELS set to study Observations the properties completed of the August most 2013 luminous galaxies GOODS-N 14’x10’ GOODS-N EGS 30’x6’ COSMOS WFC3 - 5 Orbits WFC3 - 2 Orbits 22’x8’ HUDF UDS 22’x8’ GOODS-S Grogin+ 11 Would like to use all 5 fields! 10’x13’ SameCredit: fields Ferguson covered & CANDELSwith WFC3 team Grism in AGHAST & 3D-HST Koekemoer+ 11 CANDELS- Maximizing the NumberUDS+EGS of z>6 Sources at High Luminosity +COSMOS z~7 Galaxies CANDELS- CANDELS- North+South UDS+EGS + ERS +COSMOS Ultra-Deep UDS + HUDF + Ultravista HUDF09-Ps Bowler+2014 BRIGHT FAINT UV Luminosity Bouwens+2014 Large Areas Required to Overcome Large Field-to-Field Variance Observed at High Redshift Estimated field-to- field variance for z~4-8 samples. Field-to-field variance is substantial, especially at high redshifts and at the bright end of the LF. Bouwens+2014 With lots of great data sets.... what can we learn about galaxy growth? Let’s look at the z>=4 LFs z~4-10 LFs from all CANDELS + HUDF + other legacy fields (Bouwens et al. 2014, arXiv:1403.4295, 48 pages) First Two CANDELS CANDELS CANDELS CANDELS CANDELS Frontier HUDF BORG + parallels South North UDS COSMOS EGS Fields z = 4 z = 5 Bouwens+2007 z = 6 Bouwens+2014 B14 z = 7 McLure McLure+2013 +2013 McLure z = 8 +2013 z = 10 Oesch+2014 Possible to Make Use of Whole CANDELS area? CANDELS-North/South: Deep HST data over a contiguous wavelength range Possible to Make Use of Whole CANDELS area? CANDELS-North/South: Deep HST data over a contiguous wavelength range CANDELS-UDS DEEP Y DATA 27 mag (5σ) Credit: Fontana / HUGS CANDELS-COSMOS DEEP Y DATA 26 mag (5σ) Credit: McCracken / UltraVISTA COSMOS-EGS: BUT deep ground-based data fills in wavelength gaps (most >27 mag)! z~4-10 LFs from all CANDELS + HUDF + other legacy fields (Bouwens et al. 2014, arXiv:1403.4295, 48 pages) First Two CANDELS CANDELS CANDELS CANDELS CANDELS Frontier HUDF BORG + parallels South North UDS COSMOS EGS Fields z = 4 z = 5 Bouwens+2007 Slightly More Challenging z = 6 (deep ground-based data are useful) Bouwens+2014 B14 z = 7 McLure McLure+2013 +2013 McLure z = 8 +2013 z = 10 Oesch+2014 New determinations of UV LF at z~4, 5, 6, 7, 8, 10 from all HST Legacy Fields (Bouwens et al. 2014, arXiv:1403.4295) Highlights of Bouwens+2014: > 800 likely z~7-8 galaxies > 11000 z~4-10 galaxies Provide First Determination of z~10 Luminosity Function (together with Oesch+2014) How Bright Can Galaxies Become at z~9-10? Amazingly, ~10-20x Brighter than the z~9-10 Galaxies in the HUDF Some of our best z~10 candidates are as bright as L* galaxies found at z ~ 3 by Steidel et al. 6 Oesch et al. TABLE 2 Flux Densities of z>9 LBG Candidates in the GOODS-N Field Filter GNDJ-625464314 GNDJ-722744224 GNWJ-604094296 GNDJ-652258424 B435 7 ± 93± 6 −2 ± 6 −11 ± 11 V606 2 ± 7 −3 ± 5 −5 ± 52± 8 i775 5 ± 10 6 ± 76± 7 −9 ± 11 I814 3 ± 71± 5 −2 ± 80± 9 z850 17 ± 11 −7 ± 6 −7 ± 8 −14 ± 13 Y105 −7 ± 9 −7 ± 7 −2 ± 10 −18 ± 8 J125 11 ± 812± 723± 836± 9 JH140 102 ± 47 85 ± 34 ··· 86 ± 54 H160 152 ± 10 68 ± 973± 882± 11 K137± 67 −45 ± 51 85 ± 261 76 ± 55 IRAC 3.6 µm139± 20 (< 81)* 39 ± 21 65 ± 18 IRAC 4.5 µm122± 21 (< 119)* 93 ± 21 125 ± 20 Note.—MeasurementsaregiveninnJywith1σ uncertainties. * 3σ upper limit due to uncertainties in the neighbor flux subtraction. 6 Oesch et al. 0 24 10 ric redshifts with templates that include emission lines 10 GNDJ−625464314 TABLEas 2 included in the v1.1 distribution of the code .The −2 26 Flux Densities of z>9 LBG10 Candidatesbest-fit in EAZY the GOODS-N redshifts Field are all within 0.1 of the ZEBRA AB values listed in Table 1. p(z) m −4 Filter28 GNDJ-625464314 GNDJ-72274422410 GNWJ-604094296 GNDJ-652258424 z = 10.2 (χ2 = 4.9) best 3.3. Possible Sample Contamination B435 z = 2.1 (7χ2 ±= 47.5)93± 6 −2 ± 6 −11 ± 11 low −6 V60630 2 ± 7 10−3 ± 5 As we will show−5 in± Section52 4.1, the detection± 8 of such i775 5 ± 10z 6 ± 76∼ − ± 7 −9 ± 11 GNDJ−722744224 bright z 9 10 galaxy candidates in the GOODS-N I814 3 ± 71± 5 −2 ± 80± 9 −2 dataset is surprising given previous constraints on UV z85026 17 ± 11 10−7 ± 6 −7 ± 8 −14 ± 13 LFs at z>8. A detailed analysis of possible contam- Y105AB −7 ± 9 −7 ± 7 −2 ± 10 −18 ± 8 p(z) J125m 11 ± 812−4 ± 723ination is therefore particularly± 836 important. We± 9 discuss 28 10 JH140 z = 9.9102 (χ2 =± 9.4)47 85 ± 34 ··· 86 ± 54 best several possible sources of contamination. H160 z = 2.6152 (χ2 =± 24.2)10 68 ± 973± 882± 11 low −6 K13730 ± 67 −1045 ± 513.3.1. 85 ±Emission261 Line Galaxies 76 ± 55 IRAC 3.6 µm139± 20z (< 81)* 39 ± 21 65 ± 18 GNWJ−604094296 * Strong emission line galaxies have long been known IRACOne of 4.5 Sixµ m122Bright z~9-10 Galaxies in CANDELS± 21 (<−2119) 93 ± 21 125 ± 20 26 Bright z ∼ 9 − 10 Galaxy Candidates in GOODS-North 5 10 to potentially contaminate very high-redshift sample se- NoteAB .—MeasurementsaregiveninnJywith1σ uncertainties. lection. These are a particular concern in datasets p(z) m −4 * 3σ28upper limit due to uncertainties in the neighbor flux subtrac10 tion. which do not have very deep optical data to establish a z = 9.5 (χ2 = 3.9) best z = 0.6 (χ2 = 16.5) strong spectral break through non-detections (see, e.g., low −06 3024 10 ricAtek redshifts et al. 2011; with templates van der Wel that et al. include 2011; emission Hayes et lines al. z 10 GNDJ−652258424625464314 as2012). included Sources in the with v1.1 extreme distribution rest-frame of the code optical.The line −2 26 z = 10.2 10 best-fitemission EAZY may redshifts also contaminate are all withinz ! 0.19 samples of the ZEBRA if the AB AB valuesz ∼ 10 listed candidate in Table UDFj-39546284 1. (Bouwens et al. 2011a; p(z) p(z) m m −4 Oeschet al. 2012a) is any guide. In that case, the 28 2 10 z = 9.210.2 ( χ(2χ = = 8.3) 4.9) best 3.3. Possible Sample Contamination 2 extremely deep supporting data did not result in any z = 2.22.1 (χ2 = 30.0)47.5) low −6 detectionAs we will shortward show in of Section the H 4.1,160 band, the detection but other of such evi- Fig. 2.— 6′′×6′′images30 of the four z ≥ 9galaxycandidatesidentifiedintheCANDELSGOODS-Ndata.From left to right, the images 10 show a stack of all optical bands, Y105, JH140, J125, H160,MOIRCSK,andneighbor-subtractedIRAC3.6µmand4.5µmimages.The stamps are sorted from high to lower0.5 photometric redshift1.0 from SED fits (indicated2.0 in the lower left, see4.0 also Table6.0 1). The0IRAC2.5 neighbor- 5 z7.5 10 12.5 brightdence (tentativez ∼ 9 − 10 detection galaxy candidates of an emission in the line GOODS-N at 1.6 µm subtraction works well for all sourcesGNDJ except for− GNDJ-722772274422444224, where the nearby foreground source is too bright, and clear residuals are visible atOesch+2014 the location of the candidate.