Spectroscopic Studies of z~5.7 and z~6.5 Galaxies: Implications for Reionization Spectroscopic Studies of z~5.7 and z~6.5 Galaxies: Implications for Reionization Esther Hu University of Hawaii, Institute for Astronomy IAU Colloq. 199, Shanghai 18 March 2005

Overview Scientific Motivation for Studies Scientific Motivation for Studies Ly  Emission-Line Galaxy Surveys Ly  Emission-Line Galaxy Surveys Depth, area requirements for large high-redshift samples Current status Luminosity Functions at z~5.7, z~6.5 Luminosity Functions at z~5.7, z~6.5 Line Profiles at z~5.7, z~6.5 Line Profiles at z~5.7, z~6.5 Structured Distribution of High-z Galaxies Structured Distribution of High-z Galaxies Summary Conclusions Summary Conclusions Collaborators: Len Cowie, Peter Capak, Yuko Kakazu

Lyman alpha emitters at z = 5.7 versus z = 6.5 Did the intergalactic hydrogen reionize at z just beyond 6 ? Or was it earlier? Can we find galaxies at these redshifts ? Are there enough of them to ionize the gas? What can we infer about IGM evolution from these galaxy populations? Is reionization at z=6 consistent with the galaxy observations?

Question : H REIONIZATION at z~6.2 ??? Or is it just the natural thickening of the Lyman alpha forest as we move to high redshift? VERY HARD to tell the difference between an optical depth of 10 and one of 10000! ALTERNATE TEST: is there a change in the Ly  line properties of galaxies across this redshift boundary -- either a change in the luminosity function or in the properties of the lines that might show they were now living in a mostly neutral medium where the Ly  line is cut away by the damping wings of the IGM?

High redshift galaxies We need to develop large well selected samples in the z=5 7 redshift range to answer this question…. And also-- How does initial structure formation and reionization take place? Is ionization at this redshift produced by high redshift AGN? star-forming galaxies? What is the history of star formation in the Universe in the z=5  7 range?

How can we find galaxies at these redshifts? Currently the only method which can turn up large homogeneously selected samples at these redshifts are Ly  searches with narrow-band filters or direct spectroscopic techniques (Hu et al 2002, Taniguchi et al. 2004, Ellis et al 2004) z = 6.6, ~7 Other possibilities are red color selection and targeted observations of X-ray, radio or far infrared selected sources but there are very few objects so far….

High redshift galaxies Need many very large fields Objects are sparse at brighter magnitude (few hundred per square degree for AB <25.5) Fields are highly correlated on subdegree scales Capak et al. 2004, Brodwin et al.2003 (astro- ph/ )

Hawaii survey: wide field color and narrow band mapping at z=57 Peter Capak, Esther Hu, Len Cowie, Amy Barger, Richard McMahon, Yuko Kakazu, Wei-Hao Wang, Tomoki Hayashino, Yutaka Komiyama, Ed Fomalont, Niel Brandt, Dave Alexander, Franz Bauer, Gordon Garmire, Mark Bautz, Aaron Steffen, Yuxuan Yang, Richard Mushotzky, Mauro Giavilisco, Mark Dickinson, Dan Stern, S. Okamura, C. Kretchmer, S. Miyazaki 6 Well studied fields: HDF, SSA22, Lockman Hole NW, SSA13, SSA17, A370 Deep X-ray, radio and submillimeter data for most of these Muliticolor imaging With Suprime on Subaru Megaprime on CFHT, ULBCAM on 2.2m Spectroscopy with Deimos on Keck II HK’, J, Z’, 9130/125, I, 8150/120, R, V, B, U Total area just over a square degree (Z=6.6) (Z=5.7) Spectroscopy of all X-ray and radio sources, all z=5.7 and 6.6 emission line candidates and all red color selected objects together with large magnitude selected field samples

HDF B,R,Z SUPRIME 15 / X 15 / Capak et al. 2004

HST F814W vs SUPRIME 8150 narrow band HST ‘wide-I’ continuumNB816 narrowband Wide-field narrowband exposures comparable in depth to Hubble Deep Field continuum

Filter profile 1% Night Sky Keck LRIS spectrum

z=6.56 Galaxy Behind A370 NARROW BAND (strong Ly  emission) R BAND (no galaxy detected)

N(8150) < 24 samples -- Z=5.7 selection Continuum break O II Red stars O III SSA22 Equivalent width HDF

Composite of Deimos Spectra Hu et al. (2004) R=2700 spectra allow us to easily distinguish OII and OIII emitters instrument profile

SSA22 field to N(AB)= spectroscopic Ly  emitters Spectroscopic z = 5.7 (solid boxes)

Redshift distribution of spectroscopically identified objects in Hawaii fields. 62 objects 14 objects

z=3.4 Ly  LF Z=5.7 Ly  LF (Approx 1 solar mass per year: no extinction case B) Ly  Luminosity Function at z=3.4 & z=5.7 z=3.4 Ly Sample (Cowie & Hu 1998)

Incompleteness corrected Z = 5.7 Raw Ly  selected Steidel et al,z = 3 z = 4 UV continuum luminosity function of Ly - selected objects

Z=3.4 Ly  LF Z=6.5 Ly  LF Z=5.7 Ly  LF (Approx 1 solar mass per year: no extinction case B) Lyman alpha luminosity function with z=6.5 points

Z= 5.7 Ly  emitters in SSA 22 & HDF SUPRIME fields Hu et al (no spectroscopy) HDF SSA 22

Z= 5.7 Ly  emitters in SSA 22 & HDF SSA 22 HDF Colored symbols show selected redshift regions (about 1/4 of filter bandpass)

Z= 6.5 Ly  emitters in HDF & A 370 (Hu et al. 2005) A370 and the Subaru Deep field (Taniguchi et al. 2004) are rich compared to the HDF and SSA22 (Hu et al.2004). HDF A 370 Cosmic Variance is a Problem for High-z Studies

Composite line profiles at 5.7 and 6.5 (Virtually identical!) Instrument resolution EW(5.7)=56Å EW(6.5)=50Å FWHM(5.7) =1.1Å FWHM(6.5)= 0.8Å

Does this mean conditions are the same in the IGM at 5.7 and 6.5? Maybe not:: (Madau, Haiman, Loeb, Gnedin, etc….) Maybe not:: (Madau, Haiman, Loeb, Gnedin, etc….) More luminous objects may self shield themselves by ionizing the gas around More luminous objects may self shield themselves by ionizing the gas around them. them. Even for lower luminosity objects: Even for lower luminosity objects: Clustered or neighboring objects may Clustered or neighboring objects may also ionize the region around the object. also ionize the region around the object. Preferred (low density) lines of sight may Preferred (low density) lines of sight may be ionized and we may have strong selection bias in our object sample. be ionized and we may have strong selection bias in our object sample.

H II region Neutral IGM Can we see such an emitter prior to reionization? (Haiman 2002, Madau 2002) Galaxy Maybe --- if the galaxy is bright enough IGM scattering Damping wing Intrinsic profile Final profile Small residual from red wing

Stacked profile Of 19 Ly  emitters in SSA22 Input profile Gaussian (200 km/s) truncated by IGM scattering

HST (ACS) 6000, 8000 and 9000 Å images of the z=5.7 Ly  emitters in the GOODS-N field  F606W, F775W, F850LP: only 2 of 6 are resolved by ACS 12”

Comparison of stacked colors of z=5.7 emitters with z=5.7 quasar SED of Galaxies consistent with Ly  Forest Absorption in Quasar Spectrum

Summary--- Large samples of z=5.7 and 6.5 objects can now be obtained. Ly  and continuum luminosity functions at z = 5.7 seem similar to those at lower redshifts --- (galaxies are the dominant ionizers rather than AGN) Ly  luminosity function and Ly  line shape are similar at z=6.6 and z =5.7 We may be able to make 3 dimensional maps of the cosmic web at these redshifts!

Composite line profiles of z=6.5 emitters compared with SDSS

Summary--- (ctd.) H reionization at z~6.2 ? Probably not….. Dark gaps may be a genuine Gunn-Peterson effect but could be just line blanketing of the increasing neutral hydrogen density Presence of residual flux in 1148 is strong evidence against the GP interpretation. Widths of lines would change… A370 z=6.5 emitter probably only consistent if it lies in a highly ionized hole produced by a neighbor object of by clusters of fainter objects.

Deimos Spectrum Hu et al Pre scattering spectrum (schematic). Sharp edge

SHARP EDGE Spatially resolved about 4 independent positions No sharp cutoff

GOODS-S continuum objects from Stanway, Bunker et al GOODS-N from Spinrad et al, Weyman et al., Barger et al Emission Line window

z~5.7 Ly  Emitters in GOODS-N

Stacked z~5.7 Galaxy Properties Compare the composite colors of the high-z Ly  galaxies with the spectra of bright quasars at the same redshift Compare the composite colors of the high-z Ly  galaxies with the spectra of bright quasars at the same redshift

Color-color plot: Ly  Galaxies can be distinguished from [OIII] and [OII] emitters and red objects Ly  [OIII] [OII]

Narrowband Ly  Images of z~5.7 Candidates (30”x30”)

R-band Thumbnails of z~5.7 Candidates (30”x30”)

Galaxy Colors and Quasar Spectrum

The first stars and galaxies form in the densest regions: Yellow colors show the forming stars

Searching for these first galaxies How do we find these small distant galaxies? How do we find these small distant galaxies? Can we do it from the ground or must Can we do it from the ground or must we wait for the James Webb space we wait for the James Webb space telescope (the successor to Hubble) telescope (the successor to Hubble)

YES! We can do it from Mauna Kea We can do it from Mauna Kea with its wonderful large telescopes…. with its wonderful large telescopes….

The Japanese Subaru (Pleiades) Telescope on Mauna Kea in Hawaii The largest telescope in the world with a wide-field prime-focus imaging camera Collecting area (8.3 m diameter) Collecting area (8.3 m diameter) Field of view (0.5 deg on a side) Field of view (0.5 deg on a side)

Optical Image: 50,000 galaxies in 600 square arcminutes

And much larger than the images from HST The Hubble deep field is a small part of this Subaru image which lies just above the big dipper on the sky.

Hubble Space Telescope and Subaru

Subaru can reach the depth of HST though not quite the resolution HST gives sharper images which allow HST gives sharper images which allow us to see more of the structure of the galaxies. us to see more of the structure of the galaxies.

How do we find the small early galaxies amid this sea of objects? Most of what we see in this optical image is relatively nearby. There are only a small number of very distant galaxies in the image.

Here is what a similar area of the sky might look like at a time when the universe was a tenth it present age based on simulations. The little yellow patches show the galaxies.

Theory Predictions First galaxies should be faint, because they are small First galaxies should be faint, because they are small and distant. and distant. Problem: how do you choose your candidates without examining the faint galaxies one by one? Problem: how do you choose your candidates without examining the faint galaxies one by one? POSSIBLE SOLUTIONS: POSSIBLE SOLUTIONS: Early star formation should produce strong emission from Hydrogen – easiest way to find these galaxies is to look for Lyman  (1216 Å): the strongest emission line produced by the Hydrogen atom. Early star formation should produce strong emission from Hydrogen – easiest way to find these galaxies is to look for Lyman  (1216 Å): the strongest emission line produced by the Hydrogen atom. Intergalactic gas along the line of sight to the galaxy also suppresses the spectrum at wavelengths below this line, so we can pick these galaxies out by their colours (colour selection works well at lower redshifts) Intergalactic gas along the line of sight to the galaxy also suppresses the spectrum at wavelengths below this line, so we can pick these galaxies out by their colours (colour selection works well at lower redshifts)

Lyman alpha line:

Strategy 1 Observe blank regions of sky using broadband filters and look for objects with the right colours Observe blank regions of sky using broadband filters and look for objects with the right colours Check whether the colour-selected objects are high-redshift galaxies by taking their optical spectra Check whether the colour-selected objects are high-redshift galaxies by taking their optical spectra Works well at lower redshifts but doesn’t have Works well at lower redshifts but doesn’t have the sensitivity for the galaxies at very early times. the sensitivity for the galaxies at very early times.

Strategy 2 Observe blank regions of sky using a narrow-band filter centered on the Lyman  emission at the redshift of interest Observe blank regions of sky using a narrow-band filter centered on the Lyman  emission at the redshift of interest Check whether emission-line objects detected by the filter are high-redshift galaxies by taking their optical spectra Check whether emission-line objects detected by the filter are high-redshift galaxies by taking their optical spectra

8185/105Iz’ RVB z=5.74 Galaxy (SSA22-HCM1)

Optical-Infrared SED SSA22-HCM1 Consistent with no dust reddening

z=5.74 galaxy spectroscopically confirmed with the Keck 10 metre telescopes on Mauna Kea in Hawaii

Keck LRIS spectrum Filter profile 1% Night Sky

Z=5.7 L alpha emitters in SSA22 and the HDF

We can map the filaments of the web the numerical models predict.

To get even more sensitivity we can add a second cosmic telescope to our ground telescope:

Gravitational lensing by an intervening massive cluster of galaxies will magnify the distant galaxies, making them easier to detect

This technique was used by our group here at Hawaii led by Esther Hu to find an even more distant and younger galaxy: There are now very slightly more distant objects but this is still the state of the art.

z=6.56 Galaxy behind the cluster A370 Narrow-band image (strong Ly  emission) R-band image (no galaxy detected)

Submillimetre data OPTICAL, 20cm DATA Star formation history: the formation rate of stars per unit volume of space as a function of time (1+z) 2 (1+z) 0.8 billion years 10 billion years

Future Finding high-redshift samples in large numbers: Finding high-redshift samples in large numbers: There are now about known galaxies at z=5-6 and perhaps 20 in the z=6-7 range (in the two years since the first one was found) There are now about known galaxies at z=5-6 and perhaps 20 in the z=6-7 range (in the two years since the first one was found) With large samples of these objects we should With large samples of these objects we should be able to map the filamentary structure of be able to map the filamentary structure of the universe at these redshifts. the universe at these redshifts.

Future Studying their star-formation rates: Studying their star-formation rates: Extending the star formation history to yet Extending the star formation history to yet greater distances. greater distances. Going further requires giant new infrared cameras Going further requires giant new infrared cameras which the U of Hawaii (Don Hall and Klaus Hodapp) which the U of Hawaii (Don Hall and Klaus Hodapp) are working to develop. are working to develop. Looking ahead to the James Webb Space Telescope Looking ahead to the James Webb Space Telescope

The James Webb space telescope 6 meter infrared optimized telescope. Launch about the end of the decade. Should be able To detect the smallest and most distant galaxies!

Star formation history Wilson et al and Barger, Cowie and Richards 2000 ALL SCUBA SCUBA ABOVE 6 mJy OPTICAL, 20cm DATA (1+z)^2 (1+z)^0.8