Observations of First Light

Transcription

1 Image from Space Telescope Science Institute Observations of First Light Betsy Barton (UC Irvine) Member, TMT SAC Project Scientist, IRIS on TMT Microwave Background What reionized the universe?

2 The End of the cosmic dark ages What are the open questions at high (> 7) redshifts? What is happening with present-day (or near-future) instrumentation? How do we carry it far into the future? First light science with giant telescopes like TMT What can we learn about first light in the Universe in the end?

3 Observations constraining reionization Wilkinson Microwave Anisotropy Probe (WMAP): Reionization began at 7 < z < 20 (Bennett et al. 2003; Spergel et al. 2007; depending on history) Becker et al. (2001); Fan et al. (2002): Gunn-Peterson troughs in distant quasar spectra show reionization ended near z ~ 6 Fan et al. (2001): Extrapolated quasar luminosity function not enough to reionize the universe Yan et al. (2005), Mobasher et al. (2005) and many others: Spitzer shows galaxies at z > 6 that are several hundred Myr old STAR-FORMING GALAXIES probably reionized the Universe (see Tinsley 1973!!) How can we find these galaxies?

4 Early, ionizaing seed galaxies and reionization (Iliev et al.

5 How do we guess what is out there? Theory tells us about dark matter halo sizes (small) the IGM (neutral to mixed to ionized) the very first stars (probably very big and hot) the possible progress of ionization (probably starts outside dense regions, then the bubbles coalesce) Theory tells us less about the baryonic properties of first-light sources

6 How do we guess what is out there? Extrapolate from low redshift (z < 6) Very (overly?) conservative Problem is that we are searching for the epoch when things change dramatically to metal-free star formation Try to parameterize high-z galaxies with a few parameters and see what you predict Very sensitive to what you choose Just do a thorough exploration of parameter space Eventually, it should work Expensive, and you don t know when it will work

7 Detecting sources in the continuum: often too faint for follow-up Bouwens et al. (2004) Other technique: detecting emission-line sources

8 First generation experiment: Window at z=8.227 w/ Gemini/NIRI transmission Narrow-band filter (R=125) sky emission Noise down by factor of >10 from other Gemini/NIRI narrow-band filters (Barton et al. 2004)

9 Narrow-band filter Hubble Deep Field [OII] line emission at z=2.01, 15 hours with Gemini/NIRI

10 The Production and Escape of High-redshift Lyman-α Photons partially neutral IGM (above z ~ 6.2) star formation rate stellar initial mass function { penetration through intergalactic medium escape of ionizing and Lyα photons

11 Different Lyman α scenarios yield a huge range of possibilities Scenario IMF (M O ) Metallicity (Z O ) f IGM f esc f Lyα Optimistic x Plausible x Heavy Salpeter x Salpeter x 10 41

12 z=8.2 galaxies optimistic scenario simulation 10 h - 1 comoving Mpc=5.4 arcmin (e.g., Stark & Ellis; Willis & Courbin 2005; Cubyet al. 2007; Barton et al., in prep.) 48 hour obs. 8-meter tele μm 20% throughput R=125 filter 0.35-arcsec seeing Springel & Hernquist (2003) model (Barton et al. 2004)

13 What do we know now? Stark et al. (2007) Definite sources and likely/possible candidates at z=5-10 Fluxes shifted to common luminosity distance at z=7.7

14 The Next Steps for the US: FLAMINGOS 2 and MOSFIRE Gemini-South FLAMINGOS 2: 6.1 FOV Keck MOSFIRE: 6.2 FOV DAZLE on VLT already in operation; results not publicized to date

15 F2T2 An engineering prototype for the JWST Tunable Filter Imager F2T2 will be fed by a multi-conjugate adaptive optics system and be a facility-class instrument on Gemini next year.

16 A lensed cluster survey on Gemini with AO (F2T2) z= Approx. 20 nights with F2T2 Region of parameter space where expectation value of detected sources is >= 1

17 A Single-field Deep (Unlensed) Survey z=7.7 Approx. 13 hours with Keck/MOSFIRE Region of parameter space where expectation value of detected sources is >= 1 Related science on VLT (DAZLE); planned for FLAMINGOS2

18 A Wide-field (Unlensed) Survey z=7.7 Approx. 50 hours with Keck/MOSFIRE Region of parameter space where expectation value of detected sources is >= 1 Related science on VLT (DAZLE); planned for FLAMINGOS2

19 What might we know by 2016? Difficult/impossible to explore with 8-10-m class telescopes at z > 7 z beyond ~8-10 essentially impossible

20 First-light science in the era of large telescopes TMT first-light IR instruments that will tackle first light: Thirty Meter Telescope IRMS: 2 arcminute field-of-view; can use for narrow-band imaging IRIS: 15-arcsecond fov DL imager/ arcsecond IFU

21 IRMS: InfraRed Multi-object Spectrograph Based on Keck/MOSFIRE μm All of Y, J, H, or K at once 2 arcminute field of view 46 reconfigurable cryogenic slits <= mas arcsecond sampling 80% energy in 2x2 pixels R ~ 4660

22 IRIS IFU: 64 x 64 pixel IFU; up to 128 x 128 in some modes (?) μm requirement 4 plate scales: 4, 10, 25, 50 mas 4 fields of view for IFU, at least: 0.26, 0.64, 1.60, 3.20 arcsec Full Y, J, H, or K coverage at once R ~ 4000 Design currently lenslet AND image slicer IMAGER: 15 arcsecond fov 4 mas sampling (Taylor & Prieto)

23 TMT Instruments + AO IRMS IRIS ~100 mas IQ, 2 arcmin. field Ideal to follow-up JWST surveys with spectroscopy With narrow-band filter, can pursue blind Lyman α searches Measure luminosity function, topology of reionization, maybe presence of fainter features IFU and imager Can sample best image quality produced by facility AO Mode is follow-up to IRMS or JWST In principle, best sensitivity for tiny sources (like clusters) Spatially resolved spectroscopy of larger/brighter early galaxies

24 Can we benefit from adaptive optics for first-light Lyman α science? Early galaxies are predicted to be smaller than the seeing limit (argument that IGM will scatter sources to appear big doesn t really hold water; if the IGM around them is neutral, forget about it) Stark et al. (2007) empirical best-guess scaling: mean size ~ 0.1 /(1+z/7) (right size for IRMS/coarse IRIS modes) Tiny, luminous star clusters in formation are not beyond the reach of ELTs Nominal star formation rate of unresolved 1 solar mass per year yields detectability well beyond z=10 (right size for IRIS)

25 What can we hope to understand with ELTs and AO? When did galaxies assemble and turn on? Luminosity functions to z=20 and even beyond (JWST + Lyman α searches with ELTs?) Keys are sensitivity and wide field How did reionization proceed? (Topology, rate, effects.) Maps for Mpc-scale clustering (e.g., Mesinger & Furlanetto 2007; ELTs?) Keys are sensitivity and wide field When did Population III (metal-free) stars form? Other spectral features like HeII with ELTs Key is extreme sensitivity

26 When did galaxies assemble and turn on? Baseline projects: Narrow-band filters in IRMS at range of wavelengths in J and maybe H 1-night deep exposures (single z) ~4-night wide surveys (single z) ~4-8-night cluster surveys (40 redshifts) 9-13 nights for extremely comprehensive measure of z > 7-8 Lyman alpha luminosity function

27 Changes in the ELT Era Blind surveys in the regime; lensed surveys in the regime Fluxes shifted to common luminosity distance at z=7.7

28 How did reionization proceed? Approach 1: Brute force IRMS wide-field survey for clustering of fainter sources (same survey 4-night baseline above) Clustering tells about ionized bubble sizes/fraction (e.g., Mesinger & Furlanetto 2007) (Iliev et al.)

29 How did reionization proceed? Approach 2: Find target sources in continuum (e.g., JWST legacy surveys) Search around for ionized bubbles by finding surrounding fainter Ly α sources Specific bubble mapping: 1 IRMS field ~ 5 Mpc comoving Simple surveys in

30 Topology of Reionization: Penetrating the IGM with ionized bubbles Step 1: Identify bright sources with JWST (or maybe IRMS?) Furlanetto & Oh (2005) Furlanetto, Zaldarriaga, & Hernquist (2004) And many others

31 Penetrating the IGM with ionized bubbles Step 2: Use narrow-band imaging with IRMS or IRIS to find smaller, fainter sources in surrounding ionized bubbles Best places to find PopIII Lyα may be small galaxies inside these ionized bubbles

32 When did Population III (metalfree) stars form? Luminous sources already enriched (e.g., Davé, Finlator, & Oppenheimer 2005) Step 3: Deep spectroscopy to find potential PopIII sources in surrounding ionized bubbles Best places to find PopIII Lyα may be small galaxies inside these ionized bubbles

33 HeII (1640): signature of Pop III TMT (1/2 night) Pop II Pop III Only PopIII has detectable HeII (1640) emission Schaerer (2003) predicts high (>~ 20 Angstrom) equivalent widths for young, zero-metal bursts

34 First-light Summary By 2016, z > 7 sources may become routine with meter class telescopes and wide-field OR AO imagers GSMT+AO/JWST future will include luminosity functions even to z=20; from ground, will be largely narrow-band imaging (unless OH suppression makes broad-band approaches more competitive with space) topology of reionization searches for PopIII (metal-free) star formation, or the first stars All are finite projects, range of nights with TMT first-light instrumentation Almost all have ancillary science associated with them