ARE HIGH-REDSHIFT DAMPED Ly GALAXIES LYMAN BREAK GALAXIES? 1 P. Møller. S. J. Warren. S. M. Fall. J. U. Fynbo. and P. Jakobsen - PDF

Description
The Astrophysical Journal, 574:51 58, 2002 July 20 # The American Astronomical Society. All rights reserved. Printed in U.S.A. E ARE HIGH-REDSHIFT DAMPED Ly GALAXIES LYMAN BREAK GALAXIES? 1 P. Møller

Please download to get full document.

View again

of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Data & Analytics

Publish on:

Views: 39 | Pages: 8

Extension: PDF | Download: 0

Share
Transcript
The Astrophysical Journal, 574:51 58, 2002 July 20 # The American Astronomical Society. All rights reserved. Printed in U.S.A. E ARE HIGH-REDSHIFT DAMPED Ly GALAXIES LYMAN BREAK GALAXIES? 1 P. Møller European Southern Observatory, Karl-Schwarzschild-Strasse 2, D Garching bei München, Germany; S. J. Warren Blackett Laboratory, Imperial College of Science, Technology, and Medicine, Prince Consort Road, London SW7 2BW, UK; S. M. Fall Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218; J. U. Fynbo European Southern Observatory, Karl-Schwarzschild-Strasse 2, D Garching bei München, Germany; and P. Jakobsen Astrophysics Division, European Space Research and Technology Centre, 2200 AG Noordwijk, Netherlands; Received 2001 July 16; accepted 2002 March 31 ABSTRACT We use deep HST STIS and NICMOS images of three spectroscopically confirmed galaxy counterparts of high-redshift damped Ly (DLA) absorbers (one of which is a new discovery) to test the hypothesis that high-redshift DLA galaxies are Lyman break galaxies. If this hypothesis is correct, the emission properties of DLA galaxies must lie within the range of emission properties measured for Lyman break galaxies of similar absolute magnitude. This will be true regardless of selection biases in the sample of detected DLA galaxies. We test this prediction using several emission properties: half-light radius, radial profile (Sérsic n-parameter), optical to near-infrared color, morphology, Ly emission equivalent width, and Ly emission velocity structure. In all cases the measured values for the DLA galaxies lie within the range measured for the population of Lyman break galaxies. None of the measurements is in conflict with the prediction. We conclude that the measured emission properties of the three DLA galaxies studied here are consistent with the conjecture that high-redshift DLA galaxies are Lyman break galaxies. We show that this result does not conflict with the observation that the few high-redshift DLA galaxies discovered are mostly fainter than spectroscopically confirmed L Lyman break galaxies. Subject headings: galaxies: formation galaxies: high-redshift quasars: absorption lines quasars: individual (PKS , Q , ) On-line material: color figures 1. INTRODUCTION At any redshift, the majority of the neutral hydrogen in the universe is contained in the absorbers of highest column density, the damped Ly (DLA) absorbers. Observations of the column density distribution functions f ðn; zþ for both H i and metals can be used to compute the cosmic mean densities H i and m as functions of redshift z. The metallicities of the DLA systems have a large scatter at all redshifts (Pettini et al. 1999; Prochaska & Wolfe 1999); the mean metallicity is 5% 10% solar at z 2 3 and, within the current statistical uncertainties, may or may not increase with decreasing redshift, in accordance with models of cosmic chemical evolution (Kulkarni & Fall 2002). Selection biases caused by dust can dilute the apparent rate of chemical enrichment in the DLA absorbers (Pei & Fall 1995; Boissé et al. 1998). The dust-to-metals ratio also has a large scatter among individual DLA absorbers, but the mean value 1 Based on observations made with the NASA/ESA Hubble Space Telescope and on observations collected at the European Southern Observatory, Paranal, Chile (ESO Program 63.O-0618). 51 appears to remain roughly constant with redshift at about the value in the Milky Way (Pei, Fall, & Bechtold 1991; Pettini et al. 1997). The presence of metals implies that DLA absorbers are, or have been, in a state of active star formation, and it is therefore interesting to ask whether those stellar populations are themselves visible and if they are related to other known classes of high-redshift objects. In this paper, we narrowly focus on a single question: Are highredshift DLA galaxies Lyman break galaxies? The Lyman break galaxies (LBGs; Steidel et al. 1996) are starburst galaxies at high redshift, recognizable by their relatively flat rest-frame ultraviolet continuum and a sharp discontinuity at the Lyman limit. Under the definition of LBGs we include all objects identified by their characteristic continuum shape as such on the basis of broadband photometry alone, i.e., without requiring spectroscopic confirmation. The typical brightness of an LBG is not a welldefined quantity, as it depends strongly on the flux limit of the survey. In particular, one may consider two distinct types of LBG survey; surveys concerned solely with the determination of photometric redshifts have significantly fainter flux limits than those that classify as LBGs only gal- 52 MØLLER ET AL. Vol. 574 axies with spectroscopically confirmed redshifts. It is of interest to note that to an AB magnitude limit H ¼ 26:5, i.e., flux-limited in the rest-frame optical continuum, nearly all galaxies in the Hubble Deep Field-North in the redshift range 2 z 3:5 are classified as LBGs (Papovich, Dickinson, & Ferguson 2001). With the caveat that this sample is small, it may be concluded that for observations in the optical and near-infrared, to very faint magnitudes, any galaxies with steep continuum, either due to reddening or to the presence of a population of old stars, are a minority population. So at these wavelengths, and down to the AB magnitude limit H ¼ 26:5, the term Lyman break galaxy is essentially synonymous with high-redshift galaxy. The global rate of star formation at these redshifts measured from emission from the LBGs is consistent with the rate of star formation inferred from the change with cosmic time of H i and m in the DLA absorbers (Pei, Fall, & Hauser 1999), modulo uncertain corrections in both calculations for the effects of dust. The simplest interpretation of these results is that the DLA absorbers are the reservoirs of gas from which the stars in LBGs are forming. If this is the case, for every high-redshift DLA absorber detected in the spectrum of a quasar there should be stellar emission visible from a Lyman break galaxy at the absorption redshift, coincident with or near the line of sight to the quasar. 2 Yet searches for this stellar emission have had very little success (for reviews and discussions of past surveys see Møller & Warren 1993; Kulkarni et al. 2000). The large number of unsuccessful attempts to identify the galaxy counterparts of DLA absorbers, hereafter DLA galaxies, indicates that most DLA galaxies at z 2 have very small impact parameters or are very faint (see, e.g., Steidel, Pettini, & Hamilton 1995b). Indeed, searches that have produced lists of candidates pending spectroscopic confirmation (see, e.g., Ellison et al. 2001), as well as the very limited set of confirmed identifications available at present (Fynbo, Møller, & Warren 1999), suggest that DLA galaxies mostly are fainter than L LBGs, a result supported by upper limits on H emission (Bunker et al. 1999; Kulkarni et al. 2001). This is seen by some as contradicting the view that DLA galaxies are Lyman break galaxies. In particular, the theoretical prediction that the angular momentum distribution of dark matter halos should cause LBGs and DLA galaxies to form two quite distinct populations (Mo, Mao, & White 1999) has received much attention, and it has been suggested that DLA galaxies make up a separate, low surface brightness galaxy population (Jimenez, Bowen, & Matteucci 1999). At redshifts z 0 the DLA galaxies are much easier to identify, but it is far from clear how they are related to highredshift DLA galaxies. Nevertheless, it has been found that the low-redshift DLA galaxies are also typically sub-l (Steidel et al. 1995a; Lanzetta et al. 1997; Miller, Knezek, & Bregman 1999; Cohen 2001; Turnshek et al. 2001; Bouché et al. 2001). In this paper, we present results from deep images of three spectroscopically confirmed DLA galaxies, observed with the Hubble Space Telescope (HST) instruments Space Telescope Imaging Spectrograph (STIS) and NICMOS, and make a comparison of their emission properties with those 2 The covering factor of DLA absorbers is greater than the fraction of sky covered by detectable optical emission from LBGs. So the absorbing gas would extend beyond the optically visible stellar emission. of LBGs. The images presented here are the first combined (both STIS and NICMOS) results from a large campaign of HST observations searching for 18 DLA galaxies toward 16 quasars. The targets, the strategy, and the goals are presented in detail in our first paper (Warren et al. 2001), which also presented the results of the NICMOS imaging campaign. Follow-up spectroscopy is being carried out with the ISAAC and FORS instruments on the ESO Very Large Telescope (VLT). In x 2, we present a discussion of how to use such a search for DLA galaxies to test the hypothesis that DLA galaxies are LBGs. In x 3, we describe briefly the HST observations and the measurements of the images. In x 4, we make a comparison of the measured properties, including sizes, colors, and morphologies, with the properties of LBGs of similar absolute magnitude.in x 5, we address the question posed at the top and present our conclusions. Cosmological parameters of 0 ¼ 0:3, 0 ¼ 0:7, and h ¼ H 0 =100 are assumed, and the term high redshift implies in this paper 2 z 3:5. 2. INTERPRETING SEARCHES FOR DLA GALAXIES Because of the glare from the quasar, the detectability of a DLA galaxy is dependent on both the brightness of the galaxy and the impact parameter (in addition to the brightness of the quasar and the stability of the point-spread function [PSF]). Any sample of DLA galaxies, including the small sample of three discussed here, is biased toward brighter objects and larger impact parameters. Yet if DLA galaxies are LBGs, the measured emission properties of any DLA galaxy must fall within the range of emission properties for LBGs of the same absolute magnitude. This statement is true regardless of how the DLA galaxies were selected. This is the basic prediction that we test in this paper. We compare the measured half-light radii and colors of DLA galaxies, as well as the morphologies and Ly emission properties, against the same properties measured for LBGs of similar absolute magnitude. If DLA galaxies are LBGs, they will nevertheless have a different luminosity distribution compared with a flux-limited sample of LBGs, because DLA absorbers are detected in proportion to their gas cross section. It follows that the luminosity distribution of DLA galaxies is given by the LBG luminosity function weighted by the luminosity dependence of the gas cross section. A simple calculation illustrates the importance of this for the interpretation of searches for DLA galaxies. For a luminosity function of Schechter form and a power-law relation between gas radius and luminosity R / L t, the mean luminosity of galaxies selected by gas cross section is L ¼ L ð2 þ þ 2tÞ= ð1 þ þ 2tÞ (Fynbo et al. 1999; see also Wolfe et al. 1986; Impey & Bothun 1989). For LBGs at z 3, Adelberger & Steidel (2000) measure ¼ 1:57. Depending on the value of t, the value of the mean luminosity can be much less than L. For example, for t ¼ 0:4, L ¼ 0:23L. In fact, there is some theoretical support for a value of t 0:4 (Haehnelt, Steinmetz, & Rauch 2000), based on the kinematics of DLA absorbers, as measured from the absorption profiles of low-ionization species. So the notion that DLA galaxies are typically fainter than L is perfectly compatible with the hypothesis that DLA galaxies are LBGs. We note that a consequence of this is that the value of any property that depends on galaxy luminosity will differ between galaxy samples selected by absorption No. 1, 2002 ARE HIGH-z DAMPED Ly GALAXIES LBGs? 53 (DLA galaxies) and those selected by emission (LBGs). For example, since spatial clustering is weaker for less massive (hence less luminous) galaxies (the effects of natural biasing in a hierarchical clustering scenario), we expect DLA galaxies to have a lower clustering amplitude than the emission-selected LBGs. In the longer term, when we have completed our HST survey, we will have a sample of DLA galaxies complete within well-defined selection criteria. Using the measured impact parameters, we will be able to determine the form of the relation between radius and luminosity. In fact, we have already attempted this (Fynbo et al. 1999). One point to note is that the conclusion that DLA galaxies are much fainter than L is quite robust and insensitive to the uncertainty of the faint-end slope of the LBG luminosity function. For a different value of, the value of t must be adjusted in order to reproduce the observed small impact parameters of DLA galaxies. The result is that the value of L changes little, because the exponent þ 2t appearing in the expression for L remains approximately constant. 3. PROPERTIES OF THREE DLA GALAXIES In this section, we detail the measurement of the emission properties of the three DLA galaxies, which is compared against that of the LBGs in the next section Observations and Targets The larger imaging campaign is described by Warren et al. (2001). We are obtaining high-resolution optical and near-infrared images of quasars, searching for faint galaxies close to the quasar line of sight. These are candidate DLA galaxies, targets for spectroscopic follow-up. In Warren et al. (2001) we list the 16 target quasars and their coordinates, we tabulate details of the 18 DLA absorbers, and we present details of the NICMOS observations and the list of NIC- MOS candidates. Each quasar was observed for three orbits with the NIC2 camera and the F160W filter. The NICMOS images reach to an AB magnitude of, typically, H 160 ¼ 25. A total of 41 candidate counterparts were detected in boxes of side 7 5, centered on each quasar. Spectroscopy of this candidate list is 25% complete. So far, one new, spectroscopically confirmed DLA galaxy has been discovered. The other two DLA galaxies discussed here are previous discoveries. In P. Møller et al. (2002, in preparation), we will present the candidates detected using STIS. With STIS we are imaging in the 50 CCD configuration, i.e., without a filter, in order to reach as deep as possible. In this configuration, the effective central wavelength is Å and the FWHM is Å, so 50 CCD corresponds to a very wide V magnitude. Each quasar was observed for two orbits, at different orientations, resulting in a detection limit on the combined image of, typically, V 50 ¼ 27 (AB magnitude). Because the STIS images reach much fainter magnitudes than the NICMOS images, the number of candidates in the STIS images is typically a factor of 2 3 higher than in the NICMOS images. Figure 1 reproduces the STIS images of the fields of the three quasars PKS , Q , and , in which the frames from both orbits have been registered and summed. The images show the field after subtraction of the quasar image. Strong residuals near the center of the quasar image have been masked. The DLA galaxies are indicated in each frame, labeled using the numbering scheme in x 3.2 of Warren et al. (2001). The measured impact parameters are 1 14, 0 99, and 2 51, respectively. In each case, the confirmed galaxies are the candidates in the NICMOS frames closest to the line of sight to the quasar. The three DLA galaxies have all been confirmed by spectroscopic detection of Ly in emission, but were originally found with three different search techniques. N-7-1C was identified on a deep Ly narrowband image obtained from the ground, N-14-1C was found on the image from our current STIS campaign (Fig. 1) and is also detected in the NICMOS image, and N-16-1D was found on ground-based images with the Lyman break technique. Careful PSF subtraction was in all cases instrumental in the discovery. We now briefly summarize previous results obtained on each of the three DLA galaxies N-7-1C The galaxy N-7-1C was discovered by Møller & Warren (1993), who called it S1. The DLA absorber lies at a similar redshift to the quasar. Using spectroscopic data, Warren & Møller (1996) and Møller, Warren, & Fynbo (1998) argued that the physical separation is sufficiently large that the ionizing flux from the quasar is not important. This result was supported by Ge et al. (1997), who concluded that the distance between the DLA absorber and the quasar must be larger than 1 Mpc, on the basis of ionization modeling (see also Ledoux et al. 1998; Ellison et al. 2002). Møller & Warren (1998) obtained WFPC2 images of N-7-1C and showed that the luminosity profile is similar to the profile of LBGs at similar redshift. Ge et al. (1997) reported a metallicity of the DLA absorber of 10% solar and a dust-to-gas ratio of 8% of the Milky Way value. Lu, Sargent, & Barlow (1997) were not able to fit the absorption spectrum of this DLA absorber into the rotating-disk model of Prochaska & Wolfe (1997, 1998), which requires a leading-edge asymmetry N-14-1C The galaxy N-14-1C is a new discovery. The spectrum confirming the identification is provided in Figure 2. Details are presented in Tables 1 and 2. The metallicity of the corresponding absorber is 1 3 of the solar value, but despite its being one of the most metal-rich DLA absorbers known, there is no evidence of dust (Prochaska & Wolfe 1997). In this case, Prochaska & Wolfe (1997) found evidence of the leading-edge asymmetry in the low ion absorption lines, which they interpret as evidence of a rotating disk N-16-1D The galaxy N-16-1D was first reported by Steidel et al. (1995b), who called it N1. Spectroscopic confirmation of the redshift was obtained by Djorgovski et al. (1996). Lu et al. (1997) reported a metallicity typical of DLA galaxies at such redshifts ([Fe/H] = 1.4). They also found evidence of leading-edge asymmetry in the absorption spectrum Results Photometry and Profile Fitting In Table 1, we summarize the results of STIS and NICMOS photometry and profile fitting of the galaxies. The columns list successively the quasar name, the galaxy Fig. 1. Three DLA galaxies and LBGs of similar redshift and luminosity. The three frames on the left, of side 4 00, show the images after subtraction of the quasar PSF. The DLA galaxies are marked using the numbering system from Warren et al. (2001). The top two frames are centered on the quasars, and strong residuals from the PSF subtraction have been set to the sky level near the quasar center. The object next to N-14-1C marked ps is a red point source, presumably unrelated to the DLA galaxy. The bottom left frame is centered on the DLA galaxy, and the quasar is located outside the frame. To the right of each DLA galaxy image are four frames, of side 2 00, showing LBGs in HDF-S, selected, as described in the text, to have similar redshifts and magnitudes as the corresponding DLA galaxies. [See the electronic edition of the Journal for a color version of this figure.] ARE HIGH-z DAMPED Ly GALAXIES LBGs? 55 Fig. 2. Two-dimensional spectrum of Q and of the DLA galaxy N-14-1C that confirms the galaxy as the counterpart of the DLA absorption line. Wavelength increases to the right. The bottom panel shows the summed 8000 s exposure after subtraction of sky, where the thick dark line shows the quasar spectrum and the gap represents the DLA absorption line. In the middle panel, a section of quasar spectrum centered in wavelength on the DLA absorption line has been subtracted using the SPSF software (Møller 2000). The DLA galaxy Ly emission line is visible just above the center line of the quasar spectrum and slightly redshifted (z ¼ 1:9229) relative to the absorption line (z ¼ 1:9205). In the top panel, the same data have been smoothed slightly to improve the contrast of the DLA galaxy Ly emission line with the noise. [See the electronic edition of the Journal for a color version of this figure.] name, the STIS and NICMOS AB magnitudes, the galaxy impact parameter b, i.e., the angular offset of th
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks