HII regions at high redshift Anticipating JWST NIRSpec

HII regions at high redshift Anticipating JWST NIRSpec spectroscopy Bob Fosbury, Space Telescope - European Coordinating Facility, Garching, Germany rfosbury@eso. org www. stecf. org/~rfosbury European Space Agency

March 2004, RAEF High z HII regions 2

March 2004, RAEF High z HII regions 3

Example: The Lynx arc z = 3. 357 Richness of UV spectrum High colour temperature of ionizing source(s) High ionization parameter (cf. PNe) Nebular metal abundance ~ 10 -2 Solar Stellar abundances ? ? ? Huge Lyman continuum luminosity Top-heavy IMF ? March 2004, RAEF High z HII regions 4

Time - redshift <- µwave background comes from z~1000 LBG Now Earth forms Lynx cluster z ~ 2. 5 radio galaxies Lynx arc A 1835 IR 1916, z = 10 Highest redshift quasar 3 C radio galaxies March 2004, RAEF High z HII regions 5

Outline A primary scientific driver for NIRSpec is a survey of ~2500 galaxies to measure the redshifted nebular emission lines Known as: “Kennicutt Science” Use familiar techniques to measure SFR, reddening, element (nebular) abundances, kinematics and presence of AGN Predicated on the use of the restframe optical spectrum ([OII] 3727 -> [SII] 6725) Gives Ha to z ~ 6. 6 and [OII] to z ~ 12. 4 March 2004, RAEF High z HII regions 6

Expectation is that many galaxies at high z will have emission line spectra representing the higher SFR at earlier epochs This programme is severely limited from the ground Giving Ha only to z ~ 2. 8 However, progress can be made from the ground using the restframe UV spectrum What might we expect to measure and what can be deduced about the gas in the early universe in the presence of star formation or AGN? March 2004, RAEF High z HII regions 7

Systematics of AGN spectra AGN in the UV are characterised by resonance and intercombination emission lines from regions of different density The lines result from recombination and collisional excitation processes They tell us about the nature of the ionization processes the physical condition of the gas: Te and ne the kinematics and, to some extent, the chemical composition March 2004, RAEF High z HII regions 8

Complexities include: Resonance line transfer The various effects of dust Accounting for regions of different density and continuum opacity (radiation and matterbounded clouds) Photoionization codes can be used to understand the general behaviour of the nebulosities in the UV — but it is more difficult than in the optical March 2004, RAEF High z HII regions 9

TXS 0211 -122 March 2004, RAEF z = 2. 340 High z HII regions 10

Quasar BLR Comparison of the kpcscale ISM data from radio galaxies with the Quasar BLR data discussed by Hamann & Ferland March 2004, RAEF High z HII regions 11

Spectral sequence From lowpolarization, metal-poor (? ) radio galaxies to Highpolarization, metal-rich (? ) ULIRG March 2004, RAEF High z HII regions 12

Lya/CIV & NV/CIV vs P(%) correlations March 2004, RAEF High z HII regions 13

UV spectra of AGN and Ly-break Lya NV March 2004, RAEF Si. IV OIV NIV] He. II CIV OIII] High z HII regions NIII] CIII] Si. III] 14

Systematics of Ly-break galaxies See: Shapley et al. 2003, Ap. J, 588, 65 Hot stars, HII regions, dust, outflows Relationships seen between Lya EW, continuum slope and interstellar kinematics Likely to be determined by the nature of the outflows (covering factor and velocity) and the metallicities of the HII regions March 2004, RAEF High z HII regions 15

Composite Ly-break spectra March 2004, RAEF High z HII regions 16

HII regions in the UV Local HII regions have a UV spectrum dominated by the continuum from the stellar population Emission lines are weak since the nebular excitation is quite low — ionizing O stars are typically Teff ~ 40, 000 K The presence of dust can render observations difficult e. g. , FOS spectroscopy of M 101 HII regions by Rosa & Benvenuti (1994) March 2004, RAEF High z HII regions 17

M 101 HII regions March 2004, RAEF High z HII regions 18

LHb - s relation Terlevich 2 & Melnick 2002 March 2004, RAEF High z HII regions 20

Expectations for HII galaxies at high z Low stellar metallicities ÞMassive (hot) stars => high ionization nebulæ ÞTop-heavy IMF for 1 st generation (Pop III) Low gas metallicities but get rapid pollution by Type II supernovæ The high stellar Teff means that, for a given bolometric luminosity (which is mostly below the Lyman limit) the UV/optical continuum is relatively weaker (~ T-3 eff in the Rayleigh-Jeans tail) can mean that the SED longward of Lya is dominated by the 2 -photon continuum March 2004, RAEF High z HII regions 21

Models: Panagia et al. Dependence of [OIII] on Z and Teff March 2004, RAEF High z HII regions 22

Can we observe sources like this? Look at the Lya emitters and distinguish between AGN and stellar-ionized nebulæ Search for sources dominated by emission lines => characteristic loci in colour-colour diagrams Get lucky -> the Lynx arc Magnified sources: the ‘critical line’ searches March 2004, RAEF High z HII regions 23

The Lynx arc, z = 3. 357 Discovered as part of the ROSAT Deep Cluster Survey (Holden et al. 2001) Studied in detail by Fosbury et al. (2003) Magnified by ~ x 10 by two clusters at z ~ 0. 5 March 2004, RAEF High z HII regions 24

March 2004, RAEF High z HII regions 25

UV spectrum March 2004, RAEF High z HII regions 26

Optical spectrum March 2004, RAEF High z HII regions 27

SED March 2004, RAEF High z HII regions 28

Lensing model Marco Lombardi Lensed by two clusters at z = 0. 57 and z = 0. 54 Close to a caustic in the source plane Total magnification between 8 and 16 (but could be some differential effects) March 2004, RAEF High z HII regions 29

‘Vital statistics’ Ionizing source TBB = 80, 000 ± 10, 000 K Qion = 1. 6 x 1055 ph s-1 (assuming µ(A, B) = 10) => 105– 6 massive (Pop III - like) stars Nebula ne ≤ 1000 cm-3 (from CIII]), Te ~ 20, 000 K, U = 0. 1 Z/Zsun = 0. 05, sgas ≈ 30 km s-1 (from CIII] and OIII]) March 2004, RAEF High z HII regions 30

The nebular continuum (light blue line) is simply scaled from the observed Hb flux with no reddening S 99 population models are for 107 Msun with a Salpeter IMF (a = 2. 35; 1– 100 Msun) and Z/Zsun = 0. 05 They produce ~ 20 x too few ionizing photons March 2004, RAEF High z HII regions 31

Can we find these sources in the field? Colour-redshift loci determined predominantly by the emission lines (Lynx is VERY bright at K due to [OIII]) Will appear as Lyman forest dropouts at high z March 2004, RAEF High z HII regions 32

GOODS CDF-S BViz March 2004, RAEF High z HII regions 33

GOODS ACS (V 1. 7) #6746 z. AB = 24. 43 z = 5. 6 r ~ 1 kpc Lya NIV] CIV March 2004, RAEF High z HII regions cf. Lynx F 814 WAB = 23. 1 z = 3. 36 Dm(Dz)=1. 32 34

Conclusions The very early phase of these massive starbursts is very bright We don’t see ANY starlight directly – just the HII region glow These are efficient H-ionization engines The restrame UV spectra are a lot more interesting than those of local, metalrich HII regions – the intercombination lines are good for abundance determinations March 2004, RAEF High z HII regions 35

They are considerably more massive than globular clusters We estimated ~ 109 Msun for Lynx The ionizing stars are close to our expectation for Pop III Are these the very early phase of collapse of the galaxies that produce the metals seen in the oldest globular cluster stars (~ -2 dex)? March 2004, RAEF High z HII regions 36