Neutral Hydrogen Gas in Galaxies at Moderate Redshifts: Current and Future Observations University of Cape Town 2008 Philip Lah

Collaborators: Michael Pracy (ANU) Frank Briggs (ANU) Jayaram Chengalur (NCRA) Matthew Colless (AAO) Roberto De Propris (CTIO)

Talk Outline Introduction Galaxies and Galaxy Evolution HI 21cm emission & the HI coadding technique Current Observations with the HI coadding technique HI in star forming galaxies at z = 0.24 HI in Abell 370, a galaxy cluster at z = 0.37 Future Observations with SKA pathfinders using ASKAP and WiggleZ using MeerKAT and zCOSMOS

What is HI? The many lives of hydrogen HI = neutral atomic hydrogen gas (one proton, one electron) HII = ionised hydrogen gas (one proton) – chemistry H + H 2 = hydrogen molecular gas

What is HI? The many lives of hydrogen HI = neutral atomic hydrogen gas (one proton, one electron) HII = ionised hydrogen gas (one proton) – chemistry H + H 2 = hydrogen molecular gas

The Significance of HI gas

HI Gas and Star Formation Neutral atomic hydrogen gas cloud (HI) molecular gas cloud (H 2 ) star formation

Galaxy Types

Late-Type Galaxies SpiralIrregular Usually blue in optical colour

Early-Type Galaxies EllipticalLenticular (S0) Usually red in optical colour

Late-Type Galaxy Spectrum optical spectrum of a late-type galaxy Wavelength (Å) Intensity NGC 1832 [OII] HβHβ Hα HγHγ HδHδ [OIII] [SII]

Early-Type Galaxy Spectrum line from Doyle & Drinkwater 2006 Wavelength (Å) Intensity NGC 1832 [OII] HβHβ Hα HγHγ HδHδ [OIII] [SII] optical spectrum of an early-type galaxy Wavelength (Å) Intensity NGC 1832 Mg Ca H & K G band Na

Late-Type Galaxy HI 21-cm Spectrum NGC 5701 nearly face-on spiral galaxy Radio Flux Density (mJy)

Early-Type Galaxies Little or no neutral atomic hydrogen gas As a consequence little or no active star formation

Evolution in Galaxies

Galaxy Clusters

Galaxy Cluster: Coma

Butcher-Oemler Effect

The Cosmic Star Formation Rate Density

SFRD vs z Hopkins 2004

SFRD vs time Hopkins 2004

The Cosmic Neutral Gas Density

The Cosmic Gas Density vs. Redshift Zwaan et al HIPASS HI 21cm Rao et al DLAs from MgII absorption Prochaska et al DLAs

The Cosmic Gas Density vs. Redshift Zwaan et al HIPASS HI 21cm Rao et al DLAs from MgII absorption Prochaska et al DLAs

Lyman-α Absorption Systems quasar hydrogen gas clouds Lyman-α emission Lyman-α absorption by clouds Wavelength observer Intensity

Damped Lyman-α Lyman-α 1216 Å rest frame Intensity Wavelength (Å) Lyα emission QSO redshift z = 3.2 Keck HIRES optical spectrum DLA Lyman-α forest

The Cosmic Gas Density vs. Redshift Zwaan et al HIPASS HI 21cm Rao et al DLAs from MgII absorption Prochaska et al DLAs

HI 21-cm Emission

Neutral atomic hydrogen creates 21 cm radiation proton electron

Neutral atomic hydrogen creates 21 cm radiation

photon

Neutral atomic hydrogen creates 21 cm radiation

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm emission HI 21 cm emission decay half life ~10 million years (3  s) 1 M   2.0  g  1.2  atoms of hydrogen atoms total HI gas in galaxies ~ 10 7 to M  HI 21 cm luminosity of ~2  to 2  ergs s -1 For comparison, in star forming galaxies: luminosity of H  emission ~3  to 3  ergs s -1 HI 21 cm emission ~10 7 times less power than H  emission

HI 21cm Emission at High Redshift

HI 21cm emission at z > 0.1 TelescopeRedshiftObs Time Number and HI Mass of galaxies Who and When WSRTz = 0.18 Abell hours1 galaxy 4.8  10 9 M  Zwaan et al VLAz = 0.19 Abell 2192 ~80 hours1 galaxy 7.0  10 9 M  Verheijen et al WSRTtwo clusters at z = 0.19 & z = hours 42 galaxies 5  10 9 to 4  M  Verheijen et al Areciboz = 0.17 to to 6 hours per galaxy 26 galaxies (2 to 6)  M  Catinella et al. 2007

Coadding HI signals

RA DEC Radio Data Cube Frequency HI redshift

Coadding HI signals RA DEC Radio Data Cube Frequency HI redshift positions of optical galaxies

Coadding HI signals frequency flux

Coadding HI signals frequency flux z2 z1 z3

Coadding HI signals frequency flux z2 z1 z3 velocity HI signal

Current Observations - HI coadding

Giant Metrewave Radio Telescope

Anglo-Australian Telescope

multi-object, fibre fed spectrograph 2dF/AAOmega instrument

The Fujita galaxies H  emission galaxies at z = 0.24

The Subaru Telescope

The Surprime-cam filters H  at z = 0.24

Late-Type Galaxy Spectrum optical spectrum of a late-type galaxy Wavelength (Å) Intensity NGC 1832 [OII] HβHβ Hα HγHγ HδHδ [OIII] [SII]

Intensity Narrowband Filter: Hα detection at z=0.24 AAOmega Spectrum optical red wavelengths

The Fujita Galaxies Subaru Field 24’ × 30’ narrow band imaging  Hα emission at z = 0.24 (Fujita et al. 2003, ApJL, 586, L115) 348 Fujita galaxies 121 redshifts using AAT GMRT ~48 hours on field DEC RA

SFRD vs z - Fujita Hopkins 2004 Fujita et al. 2003

Fujita galaxies - B filter Thumbnails 10’’ sq Ordered by H  luminosity

Fujita galaxies - B filter Thumbnails 10’’ sq Ordered by H  luminosity

Coadded HI Spectrum

HI spectrum all Fujita galaxies neutral hydrogen gas measurement using 121 redshifts - weighted average M HI = (2.26 ± 0.90) ×10 9 M  raw binned

The Cosmic Neutral Gas Density

my new point The Cosmic Gas Density vs. Redshift

my new point Cosmic Neutral Gas Density vs. Time

Galaxy HI mass vs Star Formation Rate

Galaxy HI Mass vs Star Formation Rate HIPASS & IRAS data z ~ 0 Doyle & Drinkwater 2006

HI Mass vs Star Formation Rate at z = 0.24 line from Doyle & Drinkwater 2006 all 121 galaxies

HI Mass vs Star Formation Rate at z = 0.24 line from Doyle & Drinkwater bright L(Hα) galaxies 42 medium L(Hα) galaxies 37 faint L(Hα) galaxies

Abell 370 a galaxy cluster at z = 0.37

Nearby Galaxy Clusters are Deficient in HI Gas

HI Deficiency in Clusters Def HI = log(M HI exp. / M HI obs) Def HI = 1 is 10% of expected HI gas expected gas estimate based on optical diameter and Hubble type interactions between galaxies and interactions with the inter-cluster medium removes the gas from galaxies Gavazzi et al. 2006

Why target moderate redshift clusters? at moderate redshifts the whole of the galaxy cluster core and its outskirts are within the field of view of a radio telescope (nearby this not the case – one has to target individual galaxies in clusters one by one) around a cluster there are many more galaxies that lie within a single telescope pointing than for a typical field the Butcher-Oemler effect – the increase in the blue fraction of galaxies in cluster cores with redshift – Is there an increase in the gas content as well?

Why target moderate redshift clusters? at moderate redshifts the whole of the galaxy cluster core and its outskirts are within the field of view of a radio telescope (nearby this not the case – one has to target individual galaxies in clusters one by one) around a cluster there are many more galaxies that lie within a single telescope pointing than for a typical field the Butcher-Oemler effect – the increase in the blue fraction of galaxies in cluster cores with redshift – Is there an increase in the gas content as well?

Why target moderate redshift clusters? at moderate redshifts the whole of the galaxy cluster core and its outskirts are within the field of view of a radio telescope (nearby this not the case – one has to target individual galaxies in clusters one by one) around a cluster there are many more galaxies that lie within a single telescope pointing than for a typical field the Butcher-Oemler effect – the increase in the blue fraction of galaxies in cluster cores with redshift – Is there an increase in the gas content as well?

Abell 370, a galaxy cluster at z = 0.37 large galaxy cluster of order same size as Coma optical imaging ANU 40 inch telescope spectroscopic follow- up with the AAT GMRT ~34 hours on cluster

Abell 370 galaxy cluster 324 galaxies 105 blue (B-V  0.57) 219 red (B-V > 0.57) Abell 370 galaxy cluster

3σ extent of X-ray gas R 200  radius at which cluster 200 times denser than the general field

The Problem of Galaxy Sizes and the GMRT

Galaxy Sizes GMRT has long baselines between dishes (up to 26 km) provides high resolution (~3 arcsec) so that the galaxies are resolved – i.e. they are not point sources for coadding the HI signal I want the galaxies to be unresolved as I can not see the size of individual galaxies for the Fujita galaxies I used an estimate of the HI size from the optical properties of spiral and irregular field galaxies and then smoothed radio data – i.e. make the galaxies unresolved

Galaxy Sizes GMRT has long baselines between dishes (up to 26 km) provides high resolution (~3 arcsec) so that the galaxies are resolved – i.e. they are not point sources for coadding the HI signal I want the galaxies to be unresolved as I can not see the size of individual galaxies for the Fujita galaxies I used an estimate of the HI size from the optical properties of spiral and irregular field galaxies and then smoothed radio data – i.e. make the galaxies unresolved

Galaxy Sizes GMRT has long baselines between dishes (up to 26 km) provides high resolution (~3 arcsec) so that the galaxies are resolved – i.e. they are not point sources for coadding the HI signal I want the galaxies to be unresolved as I can not see the size of individual galaxies for the Fujita galaxies I used an estimate of the HI size from the optical properties of spiral and irregular field galaxies and then smoothed radio data – i.e. make the galaxies unresolved

Galaxy Sizes GMRT has long baselines between dishes (up to 26 km) provides high resolution (~3 arcsec) so that the galaxies are resolved – i.e. they are not point sources for coadding the HI signal I want the galaxies to be unresolved as I can not see the size of individual galaxies for the Fujita galaxies I used an estimate of the HI size from the optical properties of spiral and irregular field galaxies and then smoothed radio data – i.e. make the galaxies unresolved

Complication The Abell 370 galaxies are a mixture of early and late types in a variety of environments. My solution make multiple measurements of the HI gas content of the coadded galaxies using a variety of resolutions

Complication The Abell 370 galaxies are a mixture of early and late types in a variety of environments. My solution is to make multiple measurements of the HI gas content of the coadded galaxies using a variety of resolutions

HI mass 324 galaxies 219 galaxies 105 galaxies 94 galaxies 168 galaxies 156 galaxies 110 galaxies 214 galaxies

HI mass 324 galaxies 219 galaxies 105 galaxies 94 galaxies 168 galaxies 156 galaxies 110 galaxies 214 galaxies

HI mass 324 galaxies 219 galaxies 105 galaxies 94 galaxies 168 galaxies 156 galaxies 110 galaxies 214 galaxies

HI mass 324 galaxies 219 galaxies 105 galaxies 94 galaxies 168 galaxies 156 galaxies 110 galaxies 214 galaxies

HI mass 324 galaxies 219 galaxies 105 galaxies 94 galaxies 168 galaxies 156 galaxies 110 galaxies 214 galaxies

HI all spectrum all Abell 370 galaxies neutral hydrogen gas measurement using 324 redshifts – large smoothing M HI = (6.6 ± 3.5) ×10 9 M 

HI blue outside x-ray gas blue galaxies outside of x-ray gas measurement of neutral hydrogen gas content using 94 redshifts – large smoothing M HI = (23.0 ± 7.7) ×10 9 M 

Comparisons with the Literature

Average HI Mass Comparisons with Coma

Abell 370 and Coma Comparison 214 galaxies 324 galaxies 110 galaxies

Abell 370 and Coma Comparison 214 galaxies 324 galaxies 110 galaxies

Abell 370 and Coma Comparison 214 galaxies 324 galaxies 110 galaxies

HI Density Comparisons

HI density field

HI density - inner regions of clusters within 2.5 Mpc of cluster centers

HI Mass to Light Ratios

HI mass to optical B band luminosity for Abell 370 galaxies Uppsala General Catalog Local Super Cluster (Roberts & Haynes 1994)

HI Mass to Light Ratios HI mass to optical B band luminosity for Abell 370 galaxies Uppsala General Catalog Local Super Cluster (Roberts & Haynes 1994)

Galaxy HI mass vs Star Formation Rate

Galaxy HI Mass vs Star Formation Rate HIPASS & IRAS data z ~ 0 Doyle & Drinkwater 2006

HI Mass vs Star Formation Rate in Abell 370 all 168 [OII] emission galaxies line from Doyle & Drinkwater 2006 Average

HI Mass vs Star Formation Rate in Abell blue [OII] emission galaxies line from Doyle & Drinkwater red [OII] emission galaxies Average

Future Observations - HI coadding with SKA Pathfinders

SKA – Square Kilometer Array final site decision by 2012?? – money will be the deciding factor both South Africa and Australia are building SKA pathfinder telescopes to strengthen their case for site selection – telescopes also do interesting science SKA promises both high sensitivity with wide field of view possible SKA sites – South Africa and Australia

SKA – Square Kilometer Array final site decision by 2012?? – money will be the deciding factor both South Africa and Australia are building SKA pathfinder telescopes to strengthen their case for site selection – telescopes also do interesting science SKA promises both high sensitivity with wide field of view possible SKA sites – South Africa and Australia

SKA – Square Kilometer Array final site decision by 2012?? – money will be the deciding factor both South Africa and Australia are building SKA pathfinder telescopes to strengthen their case for site selection – telescopes also do interesting science SKA promises both high sensitivity with wide field of view possible SKA sites – South Africa and Australia

SKA – Square Kilometer Array final site decision by 2012?? – money will be the deciding factor both South Africa and Australia are building SKA pathfinder telescopes to strengthen their case for site selection – telescopes also do interesting science SKA promises both high sensitivity with wide field of view possible SKA sites – South Africa and Australia

Why South Africa and Australia?

Global Population Density

Population Density – South Africa

Population Density – Australia

Radio Interference Frequency (Hz) Log Scales

Radio Interference Frequency (Hz) HI at z = 0.4 HI at z = 1.0 Log Scales

The SKA pathfinders

ASKAP

MeerKAT South African SKA pathfinder

ASKAP and MeerKAT parameters ASKAPMeerKAT Number of Dishes 4580 Dish Diameter 12 m Aperture Efficiency 0.8 System Temp. 35 K30 K Frequency range 700 – 1800 MHz500 – 2500 MHz Instantaneous bandwidth 300 MHz512 MHz Field of View: at 1420 MHz (z = 0) at 700 MHz (z = 1) 30 deg deg deg 2 Maximum Baseline Length 8 km10 km

ASKAP and MeerKAT parameters ASKAPMeerKAT Number of Dishes 4580 Dish Diameter 12 m Aperture Efficiency 0.8 System Temp. 35 K30 K Frequency range 700 – 1800 MHz500 – 2500 MHz Instantaneous bandwidth 300 MHz512 MHz Field of View: at 1420 MHz (z = 0) at 700 MHz (z = 1) 30 deg deg deg 2 Maximum Baseline Length 8 km10 km

ASKAP and MeerKAT parameters ASKAPMeerKAT Number of Dishes 4580 Dish Diameter 12 m Aperture Efficiency 0.8 System Temp. 35 K30 K Frequency range 700 – 1800 MHz500 – 2500 MHz Instantaneous bandwidth 300 MHz512 MHz Field of View: at 1420 MHz (z = 0) at 700 MHz (z = 1) 30 deg deg deg 2 Maximum Baseline Length 8 km10 km z = 0.4 to 1.0 in a single observation z = 0.2 to 1.0 in a single observation

single pointing assumes no evolution in the HI mass function (Johnston et al. 2007) z = 0.45 to MHz to 700 MHz one year observations (8760 hours) Simulated ASKAP HI detections

MeerKAT HI direct detections MeerKAT will detect galaxies in less time than ASKAP – due to its higher sensitivity by ~2 times – it will still take a long time to detect galaxies at z = perhaps in around a quarter of a year however at a particular redshift in a single pointing, MeerKAT will end up with fewer total detections – due to MeerKAT`s smaller field of view MeerKAT has a larger instantaneous bandwidth of 512 MHz – observe from z = 0.2 to z = 1.0 in single observation (1200 MHz to 700 MHz) MeerKAT’s field of view is better matched to many current optical and other wavelength surveys

MeerKAT HI direct detections MeerKAT will detect galaxies in less time than ASKAP – due to its higher sensitivity by ~2 times – it will still take a long time to detect galaxies at z = perhaps in around a quarter of a year however at a particular redshift in a single pointing, MeerKAT will end up with fewer total detections – due to MeerKAT`s smaller field of view MeerKAT has a larger instantaneous bandwidth of 512 MHz – observe from z = 0.2 to z = 1.0 in single observation (1200 MHz to 700 MHz) MeerKAT’s field of view is better matched to many current optical and other wavelength surveys

MeerKAT HI direct detections MeerKAT will detect galaxies in less time than ASKAP – due to its higher sensitivity by ~2 times – it will still take a long time to detect galaxies at z = perhaps in around a quarter of a year however at a particular redshift in a single pointing, MeerKAT will end up with fewer total detections – due to MeerKAT`s smaller field of view MeerKAT has a larger instantaneous bandwidth of 512 MHz – observe from z = 0.2 to z = 1.0 in single observation (1200 MHz to 700 MHz) MeerKAT’s field of view is better matched to many current optical and other wavelength surveys

MeerKAT HI direct detections MeerKAT will detect galaxies in less time than ASKAP – due to its higher sensitivity by ~2 times – it will still take a long time to detect galaxies at z = perhaps in around a quarter of a year however at a particular redshift in a single pointing, MeerKAT will end up with fewer total detections – due to MeerKAT`s smaller field of view MeerKAT has a larger instantaneous bandwidth of 512 MHz – observe from z = 0.2 to z = 1.0 in single observation (1200 MHz to 700 MHz) MeerKAT’s field of view is better matched to many current optical and other wavelength surveys

What I could do with the SKA pathfinders using optical coadding of HI if you gave them to me TODAY.

WiggleZ and zCOSMOS WiggleZzCOSMOS Instrument/TelescopeAAOmega on the AATVIMOS on the VLT Target Selection ultraviolet using the GALEX satellite optical I band I AB < 22.5 Survey Area 1000 deg 2 total 7 fields minimum size of ~100 deg 2 COSMOS field single field ~2 deg 2 Primary Redshift Range 0.5 < z < < z < 1.2 Survey Timeline2006 to to 2008 n z by survey end176,00020,000 n z in March 2008~62,000~10,000

WiggleZ and zCOSMOS WiggleZzCOSMOS Instrument/TelescopeAAOmega on the AATVIMOS on the VLT Target Selection ultraviolet using the GALEX satellite optical I band I AB < 22.5 Survey Area 1000 deg 2 total 7 fields minimum size of ~100 deg 2 COSMOS field single field ~2 deg 2 Primary Redshift Range 0.5 < z < < z < 1.2 Survey Timeline2006 to to 2008 n z by survey end176,00020,000 n z in March 2008~62,000~10,000

WiggleZ and zCOSMOS WiggleZzCOSMOS Instrument/TelescopeAAOmega on the AATVIMOS on the VLT Target Selection ultraviolet using the GALEX satellite optical I band I AB < 22.5 Survey Area 1000 deg 2 total 7 fields minimum size of ~100 deg 2 COSMOS field single field ~2 deg 2 Primary Redshift Range 0.5 < z < < z < 1.2 Survey Timeline2006 to to 2008 n z by survey end176,00020,000 n z in March 2008~62,000~10,000

WiggleZ and ASKAP

WiggleZ field data as of July 2008 z = 0.45 to 1.0 ASKAP beam size Diameter 6.2 degrees Area 30 deg 2 square ~10 degrees across

ASKAP & WiggleZ 100hrs n z = 5072

ASKAP & WiggleZ 100hrs n z = 5072

ASKAP & WiggleZ 100hrs n z = 5072

ASKAP & WiggleZ 1000hrs n z = 5072

zCOSMOS and MeerKAT

zCOSMOS field data as of March 2008 z = 0.2 to galaxies MeerKAT beam size at 1420 MHz z = 0 MeerKAT beam size at 1000 MHz z = 0.4 square ~1.3 degrees across

MeerKAT & zCOSMOS 100hrs n z = 3559 half the number of redshift

MeerKAT & zCOSMOS 100hrs n z = 3559

MeerKAT & zCOSMOS 100hrs n z = 3559

MeerKAT & zCOSMOS 1000hrs n z = 3559

HI Science with SKA Pathfinders at High z

provide constraints on the HI mass function with redshift (the distribution of galaxies with HI mass) – won’t get information on smaller HI systems – need SKA for that begin to trace how gas content varies in different environments with redshift test star formation rate – HI correlation in the period of extreme star formation activity in the universe won’t get galaxy velocity field information – again need SKA

HI Science with SKA Pathfinders at High z provide constraints on the HI mass function with redshift (the distribution of galaxies with HI mass) – won’t get information on smaller HI systems – need SKA for that begin to trace how gas content varies in different environments with redshift test star formation rate – HI correlation in the period of extreme star formation activity in the universe won’t get galaxy velocity field information – again need SKA

HI Science with SKA Pathfinders at High z provide constraints on the HI mass function with redshift (the distribution of galaxies with HI mass) – won’t get information on smaller HI systems – need SKA for that begin to trace how gas content varies in different environments with redshift test star formation rate – HI correlation in the period of extreme star formation activity in the universe won’t get galaxy velocity field information – again need SKA

HI Science with SKA Pathfinders at High z provide constraints on the HI mass function with redshift (the distribution of galaxies with HI mass) – won’t get information on smaller HI systems – need SKA for that begin to trace how gas content varies in different environments with redshift test star formation rate – HI correlation in the period of extreme star formation activity in the universe won’t get galaxy velocity field information – again need SKA

Conclusion

one can use coadding with optical redshifts to make measurement of the HI 21 cm emission from galaxies at redshifts z > 0.1 using this method we have measured the cosmic neutral gas density at z = 0.24 and have shown that the value is consistent with that from damped Lyα measurements galaxy cluster Abell 370 at z = 0.37 has significantly more gas than similar clusters at z ~ 0 the SKA pathfinders ASKAP and MeerKAT can measure HI 21 cm emission from galaxies out to z = 1.0 using the coadding technique with existing optical redshift surveys Conclusion