Lecture Thirteen: High redshift observations!

Transcription

1 Absorption-line techniques Lecture Thirteen: High redshift observations! The principal observable of an absorption line is its equivalent width W l, the fraction of light over a spectral interval that is absorbed by the gas. The principal physical parameter is the column density N, the number of atoms per unit area along the sightline. Multiplying the number of absorbing atoms per unit area by the f-value (oscillator strength) gives the number of atoms in an absorbing layer. Longair chap. 17, 18 Thursday 24 th March A few strong atomic transitions Curve of Growth This an important tool to determine N a, the number of absorbing atoms, and thus the abundances of elements in an absorbing layer, because EW varies with N a. It is a log-log plot of EW as a function of the number of absorbing atoms. Voigt profile for varying numbers of absorbing ions The shape of the line profile is also a function of the pressure which causes doppler broadening and also the global kinematics of the cloud.

2 Key Baryon Diagnostic Lines & Features Extracting Information!information about the gas from the spectral lines Question Information Observable quantity!what is it? Chemical composition Pattern of lines!what state? Molecular/atomic/ionic Pattern of lines!how hot? Temperature Widths of lines!how much? Quantity Strengths of lines!how fast? Velocity Wavelengths of lines!where is it? Location (redshift) Wavelengths of lines Prochaska & Tumlinson 2008, review arxiv: Absorption system column densities! DLAs N(HI)! 2x10 20 cm -2 mass density of neutral gas traced by DLAs, the mean metallicity of DLAs is the closest measure we have of the global degree of metal enrichment of neutral gas in the universe at a given epoch Storrie-Lombardi & Wolfe 2000 Pettini 2003 astro-ph/

3 DLAs N(HI)! 2x10 20 cm -2 DLAs N(HI)! 2x10 20 cm -2 Zn abundance as fn of redshift Wolfe et al. (1986), proposed that DLAs are the progenitors of present-day spiral galaxies, observed at a time when most of their baryonic mass was still in gaseous form. The evidence supporting this scenario, however, is mostly indirect. Prochaska & Wolfe (1998) showed that the profiles of the metal absorption lines in DLAs are consistent with the kinematics expected from large, rotating, thick disks, but others have claimed that this interpretation is not unique (Haehnelt, Steinmetz, & Rauch 1998; Ledoux et al. 1998). Protracted epoch of galaxy formation? Typically very metal poor Large range in values at same redshift Little evidence for redshift evolution Pettini 2003 astro-ph/ Pettini 2003 astro-ph/ Comparing DLAs to MW DLAs at z 2 " 3 and of stars belonging to the disk (Wyse & Gilmore 1995) and halo (Laird et al. 1988) populations in the Milky Way. # Outflows in line profiles 1406 galaxy spectra at z ~ 1.4 from the DEEP2 redshift survey The outflows have column densities of order N H ~ cm -2 and characteristic velocities of ~ km/s, with absorption seen out to 1000km/s in the most massive, highest SFR galaxies. These velocities suggest that the outflowing gas can escape into the IGM and that massive galaxies can produce cosmologically and chemically significant outflows. Both the MgII EW and the outflow velocity are larger for galaxies of higher stellar mass and SFR, with V wind ~ SFR 0.3, similar to the scaling in low redshift IR-luminous galaxies. The high frequency of outflows in the star-forming galaxy population at z ~ 1 indicates that galactic winds occur in the progenitors of massive spirals as well as those of ellipticals. Pettini 2003 astro-ph/ Weiner et al. (2009) ApJ, 692, 187

4 HI column densities Theoretical Expectations Einstein-de Sitter Universe Estimates of the baryonic mass density relative to! b for various phases of baryons in the z=3 universe. column densities corrected for obscuration Distribution fn of column density at z=2.4 Co-moving density of HI as a fn of redshift weak constraints very weak constraints Total mass density Absorption line systems traced by HI gas Uncertainties: stellar population models & IMF HI alone Metallicity in solar units Co-moving star formation rate assuming initial HI density,! HI = 4x10-3 h -1 Crude estimate to lie near stellar and atomic mass densities Pei & Fall 1995 ApJ, 454, 69 Prochaska & Tumlinson 2008, review arxiv: Baryonic mass density Evolution of HI absorbers The mass densities of stars, molecular gas, and HI are unlikely to contribute significantly to! b at z 3. Unless one invokes an exotic form of dense baryonic matter (e.g. compact objects), the remainder of baryons in the early universe must lie in a diffuse component outside of the ISM of galaxies The absence of a complete Gunn-Peterson trough in z < 6 quasars demonstrates that the majority of baryons are highly ionized at these redshifts. Because it is impossible to directly trace H+ for the vast majority of the mass in the diffuse IGM, we must probe this phase through the remaining trace amounts of HI gas

5 Evolution of Ly-" forest Modelling Why is there a break at z~1.5? How much is due to the number density evolution and how much to a possible cross-section evolution? Weymann et al ApJ, 506, 1 The evolution of N(z) is governed by two main factors: the Hubble expansion and the cosmic UV background. Coming to lower z the background starts to decrease as the number of ionising sources falls off (quasars & star forming galaxies), counteracting the Hubble expansion. Bianchi, Cristiani & Kim 2001 Metals in the IGM IGM metallicity provides information on:! History of star/galaxy formation.! Formation of unobservably early stars/galaxies.! UV ionizing background.! Feedback in galaxy formation processes. Metals in Ly! forest The lack of associated metal lines was originally one of the defining characteristics of the Ly$ forest and was interpreted as evidence for a primordial origin of the clouds (Sargent et al. 1980). However, this picture was shown to be an oversimplification by observations with sufficient sensitivity to detect the weak CIV %%1548, 1550 doublet associated with Ly$ clouds with column densities log N(H I) & 14.5 (Cowie et al. 1995; Tytler et al. 1995). Typical column density ratios in these clouds are N(C IV)/N(H I) 10 "2-10 "3, indicative of a carbon abundance of about 1/300 of the solar value, or [C/H] "2.5 with a scatter of 3 (Davé et al. 1998).

6 Where do these metals come from? STARS! Of course How Does Matter Get Out of Galaxies?! M82# A slice of the cosmic web but are these stars located in the vicinity of the Ly$ clouds observed? or are we seeing a more widespread level of metal enrichment, perhaps associated with the formation of the first stars which re-ionised the universe at z > 6? A level of metal enrichment of 10-3 to 10-2 of solar in regions of the IGM with N(H I) & cm "2 may be understood in terms of supernova driven winds from galaxies. Such outflows are observed directly in Lyman break galaxies at z = 3, and may propagate out to radii of several hundred kpc before they stall. However, if OVI is also present in Ly$ forest clouds of lower column density, as claimed by Schaye et al. (2000), an origin in pregalactic stars at much earlier epochs is probably required (Madau, Ferrara, & Rees 2001). M82# Cen & Ostriker (1999) Galaxies power strong winds that blow dust, gas, and heavy elements into the intergalactic medium. Evolution of CIV in Ly! forest Enrichment the IGM was enriched with the products of stellar nucleosynthesis from the earliest times we have been able to probe with QSO absorption line spectroscopy, only 1 Gyr after the Big Bang. The measurements of CIV suggest a metallicity Z Ly$ & 10 "3 Z! ; this is a lower limit because it assumes that the ionisation of the gas is such that the ratio C IV/C tot is near its maximum. This metallicity can in turn can be used to infer a minimum number of hydrogen ionising photons (with energy h# & 13.6 ev, corresponding to $ ' 912Å) in the IGM. Because the progenitors of the SN which produce Oxygen, for example, are the same massive stars that emit most of the (stellar) ionising photons. Assuming a solar relative abundance scale (i.e. [C/O] = 0), Madau & Shull (1996) calculated that the energy of Lyman continuum photons emitted is 0.2% of the rest-mass energy of the heavy elements produced. Songaila 2001 thus, if the Ly$ forest at z 5 had been enriched to a metallicity Z Ly$ 10 "3 Z!, then stars had emitted approximately three Lyman continuum (LyC) photons per baryon in the universe. Whether this photon production is sufficient to have reionised the IGM by these redshifts depends critically on the unknown escape fraction of LyC photons from the sites of star formation. #

7 Sizes & Metallicities Estimates of the baryonic mass density! relative to! b for various phases of baryons in the z=0 universe. The sum of the central values for the various components is significantly less than unity. Prochaska & Tumlinson 2008, review arxiv: Estimates of the baryonic mass density relative to! b for various phases of baryons in the z=3 universe. HI column density distribution function weak constraints very weak constraints Total mass density Absorption line systems traced by HI gas Uncertainties: stellar population models & IMF HI alone Crude estimate to lie near stellar and atomic mass densities Prochaska & Tumlinson 2008, review arxiv: Zwaan et al. 2005

8 HI in galaxies throughout the Hubble time! Missing Baryons... missing baryons are likely to be hidden in a warm/hot (T " 10 5 to 10 7 K) diffuse medium that precludes easy detection. Need for continuous gas accretion from the IGM cosmological simulations predict a warm/hot intergalactic medium (WHIM) comprising 30 to 40% of today s baryons, at 10 5 < T < 10 7 K The evolution of the WHIM is primarily driven by shock heating from gravitational perturbations breaking on mildly non-linear, non-equilibrium structures such as filaments. Supernova feedback energy and radiative cooling play lesser roles in its evolution. Zwaan et al. 2005, MNRAS Davé et al ApJ, 552, 473 green, over density, % ~10; red, over density, % ~10 4 Cosmic Chemical Evolution Summary infall of intergalactic material primordeal H, He cooling & collapse mixing of processes gas with ISM star formation Mass loss due to PN, stellar winds and SN nucleosynthesis in stars Mass loss due to galactic winds details of these processes are very messy and hard to model or simulate. So, simplified (semi)analytical models and assumptions are often used, e.g., the closed box model, or the instantaneous recycling approximation.

9 Observational Tools Back to more direct measurements...!... source counts & properties at high red-shift! Longair 17, 18 + literature Ellis (2006), Saas Fee lecture series, 36, First Light in the Universe, astro-ph/ surveys to unravel cosmic history, have the following tools available:! multi-wavelength studies of galaxies: tracers of structure, star formation, dust, mass assembly histories etc - richest dataset but link with theory subjective! QSOs/AGN: seem to govern feedback processes -! area of growing importance particularly for z < 3! IGM: radiation background, chemical composition - crucial for understanding reionization! Gravitational lensing: DM distribution & magnified sources - very useful for high z studies! Cosmic backgrounds: integrated contribution of unresolved sources - promising but very hard in optical/near-ir! CMB: world model and inflation: thank goodness! Evolution with Cosmic Epoch! Think global First evidence for strong evolutionary changes in the populations of extragalactic objects with cosmic epoch were found in radio source surveys in the 1950s and 1960s. An excess of faint sources was discovered in radio source & quasar surveys as compared to the expectations of uniform world models the inference was that these classes of object were much more common at earlier cosmic epochs than at present. During the 1980s the first deep optical counts of galaxies to very faint magnitudes became available thanks to CCD technology and an excess of faint blue galaxies was discovered. These studies were extended by Hubble Space Telescope This situation is repeated at all wavelengths: X-ray (ROSAT, Chandra, XMM); Infrared (IRAS, Spitzer); sub-mm (SCUBA) Longair 17, 19 + literature Ellis (2006), Saas Fee lecture series, 36, First Light in the Universe, astro-ph/

10 Source Counts for standard world models From our understanding of determining and modelling number counts and redshift distributions - there are two key relations which relate flux densities and number densities to redshift for isotropic world models Longair # 0 frequency of observation at present epoch # 1 is the emitted frequency The formula relating the distance measure, D, to redshift z for (Friedman) world models: Longair S, flux density of source is the energy recieved per unit time per unit area per unit bandwidth 1+ " number of objects in a redshift interval z to z+dz N 0 local no. density of sources, which is conserved as the Universe expands Longair r, comoving radial distance coordinate Important Implications! Observed flux density depends upon spectrum of source because radiation emitted at # 1 is observed at red-shifted frequency # 0. Differences between inverse square law and predictions of standard world models taking into account different region of spectrum by K-correction! In standard world models the distance tends to finite limit as z ' # Except in pathological cases (e.g. dust in sub-mm) the effect of observing the source spectrum at redshifted frequency generally results in the sources becoming fainter with increasing redshift! Volume element per unit redshift changes from dn(z) ( z 2 dz for z <<1 to dn(z) ( z -3/2 dz for! 0 z >> 1. This means that the volume elements become smaller and smaller with increasing redshift resulting in a cut-off to the source distribution at redshifts! 0 z >> 1 Progressively more and more difficult to discover high red-shift objects; discrimination in both flux and observable volume numerical integration required! & = 0 D, distance measure Example: Population of sources which have power law spectra, L(#) ( # -", which is good approx for extra-galactic radio sources, X-ray sources and quasars differentiating Longair For a uniform population of sources of local space density N 0 (L) : Predicted source counts Normalised differential source counts for a single luminosity class of source having a spectral index "=1 for different values of density parameter! 0. Depart rapidly from Euclidean expectations even at low z! & = 0! 0 = Different slope in the integrated source counts already at z=0.5 For a locally Euclidean population, N(>S) ( S -3/2 2.3 Longair! 0 +! & = 1! 0 = Einstein-de Sitter,! 0 =1,! & = 0 A uniform simple population must have slope smaller than Euclidean predictions. The fact that X-gal radio sources have much steeper slope than expected is strong evidence that these sources are more common at large redshifts - which means strong evolution in source properties (also GRBs).

11 Passive Evolution of Giant Elliptical Galaxy in its rest frame. Optical Sources Counts depend strongly on SFHs if observing UV! passive evolution Passive Evolution Most IR luminosity of a galaxy originates from stars in the RGB and the colours and luminosities are to first order independent of the mass of the MS progenitor. Mass-luminosity relation for stars on MS: lifetime of RGB much shorter than MS MS lifetime determined by time it takes star to burn 10% of its mass into He (Schönberg-Chandrasekhar limit) z= During MS evolution the luminosity of stars are constant, and since the available fuel is directly proportional to the mass of the star, the MS lifetime, t: t!, is the MS lifetime of the Sun ~10 10 yrs. rest frame wavelength for observations in K for different redshift galaxies Spectral evolution model (Bruzual & Charlot 2003) of a giant elliptical galaxy in which the stellar population formed in an initial starburst which lasted ~10 8 yrs For M~M!, x=5 this means for M=2M! t =1/16 t! = 6x10 8 yrs. range of masses that contribute to the light of old stellar populations is 1 < M! < 2 Tinsley & Gunn (1976) ApJ, 206, 525 Mass fraction per log mass bin IMF < 1 M! makes minor contribution to light but is very important for mass inventory Initial Mass Function Salpeter > 1 M! reproduces colors and H" properties of spirals Passive Evolution (contd.) Assuming all stars formed in an initial brief period & subsequent luminosity evolution is just due to stellar evolution of the population: IMF: for Salpeter, y= 2.35 no of RGB stars using IMF SFR the relation between mass and MS lifetime x=5, y= Mass (M! ) for a critical world model, t/t 0 = (1+z) -3/2 and so At z=1 this stellar pop is 2x more luminous than at present and at z=3 it will be 4x as luminous.

12 HUDF Galaxy Counts HDF Spectroscopic follow-up is challenging Importance of High z Data!Our understanding of the local population of galaxies is largely confined to its spatial distribution and (perhaps) some environmental trends in the broadest sense! Standard model needs major `additional ingredients to reconcile local color and luminosity distributions: such effects have major implications for assembly and SF histories!three key 1 < z < 4 galaxy populations whose studies are relevant to addressing these issues Metcalfe et al. 1996, Nature, 383, 263 I: Lyman break galaxies: color-selected luminous star forming galaxies z > 2 II: Sub-mm galaxies located via redshifted dust emission Little change in K no counts to z~1 which can be interpreted as the old population being in place by then. III: Various categories of passively-evolving sources whose selection has been enabled via deep IR data Abraham et al. 1996, MNRAS, 279, L47

13 Finding star-forming galaxies at high z Photometric Cuts: Prediction and Practice The Lyman continuum discontinuity is particularly powerful for isolating starforming high redshift galaxies. From the ground, access to the redshift range z=2.5-6 in the micron range. Steidel et al 1999 ApJ, 462, L17 Steidel et al 1999 ApJ, 519, 1 Steidel et al 2003 ApJ, Expectations Real Data (10 field) Spectral energy distributions allow prediction where distant SF galaxies lie in color-color diagrams such as (U-G vs G-R) (Steidel et al 1996) Spectroscopic Confirmation LBGs HST images of spectroscopically confirmed Lyman break galaxies with z>2 in HDF-N revealing small physical scale-lengths and irregular morphologies Giavalisco et al 1996 ApJ 470, 189

14 V-band Luminosity Function at z"3 Local LF Studying LBGs Key to physical nature of LBGs is origin of intense SF: -prolonged due to formation at z~3 (Baugh et al 1998)? or -! temporary due to merger-induced star burst (Somerville et al 2001)? Shapley et al 2001 Ap J 562, 95 Shapley et al ApJ LBG Properties (z"3) <age> = 320 z = 3 <M*> = ~2 x M! FIRES VLT+ISAAC 176 hours J,H,K imaging <E(B-V)> =0.15 A UV ~1.7 )~5 <SFR> ~ 45 M! yr -1 Extinction correlates with age young galaxies are much dustier SFR for youngest galaxies average 275 M! yr -1 ; oldest average 30 M! yr -1 Objects with the highest SFRs are the dustiest objects Shapley et al. (2001) ApJ, 562, 95 MS N. Forster-Schreiber et al. 5.4 x 5.4, seeing 0 45 HDF South I. Labbe et al. 2.4 x 2.4, seeing 0 45

15 van Dokkum et al. 2004, 2006 Z=2.43 Z= Z=2.43 Surprisingly high surface density: ~0.8/arcmin 2 to K=21 (two fields) ~2/arcmin 2 to K=22 (HDF-S) ~3/arcmin 2 to K=23 (HDF-S) Z=2.71 Z=3.52 van Dokkum,Franx, Rix et al IR selected galaxies (DRGs) Distant Red Galaxies LBG Summary!Period of elevated star formation (~100 s M! yr -1 ) for ~50 Myr with large dust opacity (sub-mm galaxy overlap)!superwinds drive out both gas and dust, resulting in more quiescent star formation (10 s M! yr -1 ) and smaller UV extinction!quiescent star formation phase lasts for at least a few hundred Myr; by end at least a few M! of stars have formed!all phases are observable because of near-constant far-uv luminosity van Dokkum et al. 2004