A Groundbased Imaging Study of Galaxies Causing DLA, subDLA, and
LLS Absorption in Quasar Spectra
We present results from a search for galaxies that give rise to damped Lyman alpha (DLA), subDLA, and Lyman limit system (LLS) absorption at redshifts in the spectra of background quasars. The sample was formed from a larger sample of strong Mg ii absorbers ( Å) whose H i column densities were determined by measuring the Ly line in HST UV spectra. Photometric redshifts, galaxy colours, and proximity to the quasar sightline, in decreasing order of importance, were used to identify galaxies responsible for the absorption. Our sample includes 80 absorption systems for which the absorbing galaxies have been identified, of which 54 are presented here for the first time. In some cases a reasonable identification for the absorbing galaxy could not be made.
The main results of this study are: (i) the surface density of galaxies falls off exponentially with increasing impact parameter, , from the quasar sightline relative to a constant background of galaxies, with an e-folding length of kpc. Galaxies with kpc calculated at the absorption redshift are statistically consistent with being unrelated to the absorption system, and are either background or foreground galaxies. (ii) is inversely correlated with at the 3.0 level of significance. DLA galaxies are found systematically closer to the quasar sightline, by a factor of two, than are galaxies which give rise to subDLAs or LLSs. The median impact parameter is 17.4 kpc for the DLA galaxy sample, 33.3 kpc for the subDLA sample, and 36.4 kpc for the LLS sample. We also find that the decline in with can be roughly described by an exponential with an e-folding length of 12 kpc that occurs at . (iii) Absorber galaxy luminosity relative to , , is not significantly correlated with , , or . (iv) DLA, subDLA, and LLS galaxies comprise a mix of spectral types, but are inferred to be predominantly late type galaxies based on their spectral energy distributions. (v) The properties of low-redshift DLAs and subDLAs are very different in comparison to the properties of gas-rich galaxies at the present epoch. A significantly higher fraction of low-redshift absorbers have large values, and a significantly higher fraction of the large value galaxies have luminosities . The implications of these results are discussed.
keywords:quasars: absorption lines – galaxies: ISM – galaxies: statistics
The recognition that damped Ly absorption-line systems (DLAs)
seen in quasar spectra arise in neutral-gas-rich foreground galaxies
(Wolfe et al. 1986) motivated new methods for high-redshift galaxy
studies years ago. These high HI column density systems,
with atoms cm, trace the bulk of
the observed neutral gas in the Universe, and they are, therefore,
powerful probes of galaxy formation and evolution back to the
redshifts of the most distant quasars. Larger datasets and deeper
surveys (e.g. Prochaska, Herbert-Fort, & Wolfe 2005; Rao, Turnshek,
& Nestor 2006, henceforth, RTN06; Noterdaeme et al. 2009) have
improved our knowledge of the neutral gas content and distribution at
all observable redshifts, including the present epoch (Ryan-Weber et
al. 2003; Zwaan et al. 2005). No other technique has revealed
comparable reservoirs of neutral gas beyond the local Universe. See
Rao (2005) and Wolfe et al. (2005) for some past reviews. Few DLAs
were known at low redshift prior to the turn of the century because
the Ly line falls in the UV for . Thus, in the absence
of large UV spectroscopic surveys, this meant that studies of neutral
gas in what corresponds to the most recent % of the age of
With the identification of DLAs (and subDLAs and LLSs), follow-up work
involving the study of their host galaxies, environments,
neutral-gas-phase metallicities, kinematics, 21 cm spin temperatures
(when possible), ionization conditions, and numerical and
semi-analytic modeling has kept many astronomers busy for
Semi-analytic and numerical models, some of which are based on results from high-resolution spectroscopic data of DLA metal absorption lines, have resulted in a variety of often competing scenarios for DLAs: large rapidly rotating protogalactic disks (Prochaska & Wolfe 1997, 1998; Wolfe & Prochaska 1998), merging protogalactic clumps in a hierarchical merging scenario (Haehnelt et al. 1998), low surface brightness galaxies (Jimenez et al. 1999), dwarf galaxies (Okoshi & Nagashima 2005), compact, faint galaxies with impact parameters smaller than 5 kpc at (Nagamine et al. 2007), and the outer regions of high- Lyman break galaxies (Møller et al. 2002; Wolfe et al. 2003). Recent high-quality H i 21 cm data of local galaxies indicate that DLA gas velocity widths are more consistent with tidal gas related to galaxy interactions or superwinds rather than galaxy disks (Zwaan et al. 2008). In addition, some recent compelling cosmological simulations relevant to interpreting the nature of Mg ii -selected galaxies in general, and DLA galaxies in particular, have been presented by Kacprzak et al. (2010).
However, there is a dearth of identified DLA (and subDLA) galaxies, and this has undoubtedly motivated the various interpretive scenarios. Therefore, larger samples of neutral-gas-selected galaxies are required to investigate the possibilities, which in turn will help constrain models of galaxy evolution and better establish the galaxy population that harbors the bulk of the neutral gas in the Universe. Traditional galaxy surveys trace galaxies by virtue of their luminous emission. Beyond the local Universe far less is known about neutral-gas-selected galaxies and their relationship to luminosity-selected galaxies.
As an extension of our earlier work (Rao et al. 2003), we have undertaken a large multi-colour optical/IR imaging programme of quasar fields containing Mg ii absorbers with measured . The absorbers were selected from Table 1 of RTN06, and were required to have absorption redshifts to optimize the possibilities for detection and characterization of galaxies in these fields. Data and results from eight DLAs in six quasar fields were presented in Turnshek et al. (2001), Rao et al. (2003), and Turnshek et al. (2004). Other observations from the literature were also considered. In total, 27 DLA, 30 subDLA, and 23 LLS galaxies (i.e. 80 absorbers in total) have been identified. In this paper, we present and analyse the entire dataset. These observations are described in §2. The identification of galaxies is described in §3 through specific examples. The entire dataset can be accessed on line at http://enki.phyast.pitt.edu/Imaging.php. A summary of the data and statistical inferences are presented in §4. Conclusions and discussion are presented in §5 and 6, respectively. Among other findings this work demonstrates that the neutral hydrogen column density, , is strongly correlated with impact parameter, , in the sense that DLA galaxies are systematically closer to the quasar sightline, by a factor of two, than are galaxies which give rise to subDLAs and LLSs. We also find that the properties of low-redshift () DLAs and subDLAs are very different in comparison to the properties of H i -rich galaxies at the present epoch. A significantly higher fraction of low-redshift absorbers have large , and a significantly higher fraction of the large galaxies have luminosities .
All magnitudes reported in this paper are in the AB system, and all distance related quantities are calculated using the “737” cosmology with () = (0.7,0.3,0.7).
The imaging data were obtained between December 1998 and June 2005 through community-access time at national facilities as well as through Ohio State University’s share of time at the MDM Observatory. The various telescopes and detectors that were employed, as well as the varying observing conditions that prevailed over the better part of a decade of observations, resulted in an unavoidably inhomogeneous dataset. Nevertheless, since most of the data are well calibrated and reach fainter magnitudes than large groundbased surveys such as the Sloan Digital Sky Survey, this observing programme has yielded the most useful and comprehensive set of images of DLA, subDLA, and LLS absorption-line-producing galaxies that has thus far been obtained.
The optical images were obtained at Kitt Peak National Observatory in Arizona. The telescopes and corresponding detectors used were 1) the KPNO 2.1 m with the T2KA or T2KB CCDs covering a 10.4′ 10.4′ field-of-view at a scale of 0.305/pixel, 2) the MDM Observatory 2.4 m Hiltner with the 1024 1024 Templeton CCD covering a 4.72′ 4.72′ field-of-view at 0.275 /pixel, and 3) the 3.5 m WIYN with the Tip-Tilt Module (WTTM) covering a 3.84′ 4.69′ field-of-view at 0.1125/pixel. The near-infrared images were obtained on Mauna Kea, Hawaii, with the NASA IRTF 3.0 m telescope using NSFCAM which has a 76.8 76.8 field-of-view at 0.30 /pixel. The detector was the 256256 InSb array. A few images were obtained with the SpeX infrared slit-viewer/guider covering a 60 60 field-of-view at 0.12 /pixel with the Raytheon 512 512 InSb array. The optical images taken with the KPNO 2.1 m or WIYN telescopes were obtained using the Johnson-Cousins or KPNO SLOAN filters, and those observed at MDM were observed with the MDM Gunn-Thuan filters. Henceforth, we ignore the differences between the SDSS and Gunn-Thuan filter sets since the transformation between them is small. Near-infrared images were taken using the standard Mauna Kea Observatory filter set.
Optical data were taken in groups of 3 or 4 offset exposures ranging from 900 to 1800 seconds per exposure, and standard data reduction procedures were followed. The infrared observations were carried out using a series of either 30 or 60 dithered short exposures ranging from 2 to 20 seconds per exposures. The individual exposure times were chosen to prevent the quasar point spread function from saturating. Flat fielding was done using sky frames constructed from the dithered object frames. Landolt standards were used to calibrate the Johnson-Cousins observations (Landolt 1992). Photometric calibration of fields that overlapped with SDSS images was performed by comparing our instrumental magnitudes with the SDSS DR4 photometry of point sources. The photometric zeropoint solution with corresponding errors for each frame were determined by a least squares fit to the SDSS magnitudes and our instrumental magnitudes. Generally, 10 or more isolated, unsaturated stars that were common to our images and the SDSS fields were used in the calibration. UKIRT faint photometric standards (Hawarden et al. 2001) were used to calibrate the near-infrared observations.
Table 1 lists the fields for which we obtained imaging data. The first six columns give details about the absorption-line system from RTN06. Column 1 gives the quasar name, column 2, the quasar magnitude, column 3, the quasar emission redshift, column 4 gives the redshift of the Mg ii absorption-line system, column 5 the rest equivalent width of the stronger member (2796) of the Mg ii doublet, and column 6 gives the H i column density as measured from the HST UV spectrum. Column 7 lists the optical and infrared filters through which images of each quasar field were obtained. Here we summarize a few salient features of the dataset.
We were able to obtain a complete optical and infrared (UBRIJHK or ugriJHK) dataset for 18 of the 60 fields. The largest number of fields, 53, were observed in K, while the fewest, 26, were observed in U (or u). Figure 1 gives the distribution of 3 surface brightness limits reached in each filter. The infrared data show the least spread since almost all were obtained with the IRTF NSFCAM, with the exception of two observations with the (low-sensitivity) IRTF SpeX guide camera.
Since the K-band data are the most extensive as well as uniform, we illustrate a few properties of the dataset using the K-band data sample. Figure 2 is a plot of the K-band galaxy luminosity limit (in terms of , where for , Cirasuolo et al. 2006) as a function of redshift. Here, the limiting K-band surface brightness was used to estimate the luminosity of a fiducial 10 kpc-sized galaxy that can be detected at the redshift of the absorber. K-corrections appropriate for an Sb type galaxy have been applied to all datapoints.
The distribution of seeing values for our data, expressed in terms of the FWHM of a point source, is shown in Figure 3. The seeing is a particularly important parameter for this study because the quasar point spread function (PSF) limits our ability to study the smallest impact parameters. For ground-based imaging, techniques such as adaptive optics (AO) achieve seeing values down to a few tenths of an arcsecond. However, the presence of a bright (12 to 15 magnitude), nearby (within 30″), point source is generally required for implementing AO techniques. Since the (faint) quasar is the brightest object in the majority of our fields, we could not take advantage of AO. From Figure 3 it can be seen that the optical dataset has seeing values generally 1″, while for the infrared data, seeing values 1″ were often achieved. We were able to probe smaller impact parameters than the seeing radius for images where the quasar PSF could be subtracted. However, this was not possible for all fields, because suitable PSF stars were not always available within the image, and we did not observe PSF stars separately. In Figure 4 we plot minimum detection impact parameter histograms for the K-band images in arcsec as well as in kpc at the redshift of the absorbers. The minimum detection impact parameter for fields where the quasar PSF could not be subtracted is conservatively taken to be the radius at which the PSF blended with the background. This estimate depends on the brightness of the quasar as well as on the seeing. For PSF-subtracted fields, the minimum detection impact parameter is measured as the radius of the mask that was applied to the subtraction residuals. The average minimum impact parameters for all fields is kpc. The average minimum impact parameter for PSF-subtracted fields is kpc, and for non-PSF subtracted fields, it is kpc. We note here that whether the quasar PSF was, or was not, subtracted in our ground-based data has not influenced the identification of absorber galaxies (§3 and §4). For our -band sample, we find that the distribution of impact parameters of absorber galaxies identified in our PSF-subtracted fields and non-PSF-subtracted fields are similar, with Kolmogorov-Smirnov (KS) test probability .
The identification of galaxies causing absorption in quasar spectra is, undoubtedly, best achieved from space. Not being able to probe to within 5 or 10 kpc of the quasar sightline is the most severe limitation of a groundbased imaging programme. Low luminosity dwarf galaxies directly along the quasar sightline will most likely be missed, resulting in an incorrect identification of the absorbing galaxy. Below, we attempt to quantify the possible number of missed galaxies in our sample due to this bias.
3 Identification of Absorbing Galaxy Candidates
The detection and photometry of sources were carried out using the automated software SExtractor (Bertin & Arnouts 1996). The SExtractor input parameter that defines the detection threshold for source identification was set to 1 above the sky background, and the minimum detection area was set to 5 adjoining pixels. “Adjoining” as implemented in SExtractor refers to any pixels touching at corners or sides. A source is considered to be a confident detection if it was detected at the or higher level through more than one filter. Its position was determined using the image with the best seeing.
Figure 1 shows that most of the -band data (38 fields) reach surface brightnesses between 21.5 and 22.5 magnitudes per square arcsec at the 3 level. Figure 5 gives the redshift distribution of the 38 absorbers in these fields, 24 of which have redshifts . This is a small enough redshift interval that we use a single value for the angular diameter distance to estimate the surface density of galaxies. We then use this sample to estimate: (1) the background (and foreground) number density of galaxies, (2) the excess around the quasar line of sight that can be attributed to the presence of an absorbing galaxy or galaxies associated with it, and (3) the number of absorbing galaxies that might have been missed due to the glare of the quasar PSF. Figure 6 shows the number of galaxies per square kpc as a function of impact parameter from the quasar calculated in annuli of width 10 kpc. The red line is the best-fit exponential profile to the data with an e-folding length of 46.1 kpc. The horizontal asymptote, which occurs at kpc, is shown by the blue dotted line. It represents the background plus foreground galaxy number density. Galaxies beyond kpc can be considered background or foreground galaxies that are not associated with the absorption systems. Moreover, assuming that the distribution can be extrapolated to impact parameter gives an estimate of the number of galaxies unaccounted for due to the presence of the quasar PSF and the inability to subtract it perfectly. This suggests that the expected number density at kpc is galaxies per square kpc per quasar field, and that galaxy candidates per field might have been missed. Or, on average, one in every 14 fields may have a galaxy at kpc that is not identify in our groundbased imaging survey. This amounts to approximately four among the 55 identified candidate galaxies in our survey (§4).
Based on the above analysis, only objects within an impact parameter kpc from the quasar at the absorber redshift (or lowest absorber redshift in the case of multiple absorbers per quasar sightline) are catalogued for each field, since galaxies farther away can statistically be considered background or foreground galaxies.
The absorbing galaxy has not been confirmed spectroscopically for any
of the fields presented here. A spectroscopic redshift that matches
the absorption redshift would, of course, lead to a more confident
identification of the absorbing galaxy (or a parcel of gas associated
with it). In the absence of spectroscopic data, we assign a galaxy as
a “candidate absorber” with varying levels of confidence based on
several criteria. The highest level of confidence is achieved when a
galaxy’s photometric redshift matches that of the Mg ii absorption-line system within the uncertainties. Photometric redshifts
were determined for galaxies that were detected in four or more
We assign a “CL” value, or confidence level, for each galaxy identification. Galaxies which have been confidently identified through photometric redshifts that match the absorption redshift are labeled as having CL = 1. Identifications which were made based on colours that were consistent with a galaxy being at the absorption redshift, the proximity criterion, or photometric-redshift matches that were only marginally consistent with the absorption redshift are assigned confidence level CL = 2 or 3, with 2 being the more confident identification. No galaxy identification was possible for a few fields. For example, this may happen if an absorber redshift does not match the photometric redshift of any of the galaxies in a given field, or when galaxy colours are ambiguous or are consistent with a large redshift range. These fields are not assigned a CL value.
We now illustrate our process of absorbing galaxy identification with a few representative examples that include most of the issues we faced while assigning galaxies to absorption systems. The images and photometry for all objects in our sample are available on line at http://enki.phyast.pitt.edu/Imaging.php. We also provide results from photometric redshift and stellar population synthesis template fits, details of which are explained in the Appendix. Readers who are not interested in the details of galaxy selection can skip to §4.
Here we provide our reduced images, photometry tables, and photometric redshift fits and derived stellar population synthesis parameters for four fields. Sources detected within 100 kpc of the quasar sightline at the absorption redshift (or smallest absorption redshift for multiple absorbers along the same sightline) are numbered in order of increasing impact parameter from the quasar, and ellipses are drawn around sources in each image only to guide the eye. Photometry tables give positions relative to the quasar, AB magnitudes, and the detection significance, “DS”, which is defined as the number of standard deviations the source is detected above the background. , where is the net source counts, is the counts per pixel that correspond to a source detected at 1 above the background, and is the number of pixels within the detection isophote. A source is considered to be a detection if and . Tables describing photometric redshift fits give details of the stellar population templates that best fit the photometry. The information provided includes object number as marked on the images and its projected distance from the quasar in arcsec and kpc assuming that the galaxy is at the absorption redshift, age of the stellar population, star formation rate e-folding time, , extinction, , metal mass fraction, (), and the photometric redshift and error.
Example 1: the 0153+0009 field
This is an example of our highest level of confidence for absorbing galaxy identification, where the photometric redshift of the galaxy with the smallest impact parameter to the quasar matches the redshift of the absorption-line system.
The sightline towards the quasar 0153+0009 (SDSS J015318.19+000911.3) contains a subDLA system at with a column density of (RTN06). We obtained and images of this field (see Figure 7). We have used SDSS optical photometry for this field to supplement our infrared data; together the data were used to determine its photometric redshift. Our measured photometry is given in Table 2. The quasar PSF could not be subtracted as there were no suitable PSFs stars in the field, and so the quasar has been masked out. We detect nine objects within 100 kpc of the quasar.
Object 1 is at = 4.9, which is equivalent to 36.6 kpc at the absorber redshift. SDSS photometry for Object 1 was obtained from S. Zibetti (private communication) since it is not in the “photoObj” catalogue made available in the SDSS database. S. Zibetti ran his PSF subtraction software on the SDSS image (Zibetti et al. 2007), and provided us with the photometry of Object 1 in all five SDSS bands (Table 3). A photometric redshift of = 0.7450.040 is derived for Object 1 by supplementing these magnitudes with our infrared data (see Table 4). The stellar population template fit is shown in Figure 8. This is consistent with the absorption redshift to within the errors, and therefore, Object 1 is considered to be the absorbing galaxy.
The = 1.38 colour of Object 2 is not consistent with it being at the absorption redshift. Object 3 is included in the photometry table because it looks real by eye. However, based on its detection significance, DS, we do not consider it to be a confident detection. No redshift information could be extracted from the IR data on Objects 4, 5, and 8; they are not detected in the SDSS images.
Object 6 is identified as a star in the SDSS database,
however, it is extended in our images. Objects 7 and 9 have , and respectively, according to
the SDSS database
In summary, due to its matching photometric redshift as well as proximity, Object 1 is selected as the absorbing galaxy with CL = 1. The photometric redshift derived for Object 6 is consistent with the absorption redshift, making it likely that Objects 1 and 6 are members of the same galaxy cluster or group.
|QSO||0.0||0.0||0.0||17.18 0.003||31.9 (188)||17.58 0.01||16.8 (120)||17.45 0.004||20.5 (138)|
|1||4.7||1.4||4.9||21.43 0.07||2.9 (41)||21.39 0.10||2.7 (22)||21.01 0.07||3.0 (35)|
|2||+2.5||+4.7||5.3||21.88 0.09||2.3 (34)||21.82 0.13||1.9 (21)||20.50 0.05||2.7 (63)|
|3||6.2||+3.4||7.1||23.22 0.22||1.7 (8)|
|4||+3.6||6.5||7.4||22.10 0.08||2.9 (22)||21.55 0.11||2.6 (20)||21.19 0.07||3.5 (26)|
|5||1.7||+7.7||7.9||23.09 0.16||2.0 (13)||22.95 0.21||2.0 (9)|
|6||0.4||+10.8||10.8||21.90 0.08||3.0 (26)||22.00 0.12||2.9 (12)||21.58 0.08||3.3 (19)|
|7||+10.8||+2.6||11.1||22.20 0.09||2.9 (20)||22.54 0.16||2.6 (8)||21.85 0.10||2.6 (19)|
|8||8.1||+7.6||11.1||23.12 0.20||2.2 (7)|
|9||7.3||+10.0||12.4||22.09 0.09||2.6 (25)||23.10 0.21||2.1 (6)||22.23 0.13||2.2 (16)|
|Galaxy||Stellar Population Synthesis Model Parameters|
Example 2: the 0735+178 field
This is an example of a field where the three objects with the smallest impact parameters are ruled out as absorbing galaxy candidates. The fourth closest object is the best candidate for the absorbing galaxy.
The sightline towards the quasar 0735+178 contains a LLS at = 0.4240 with a column density (RTN06). This is an interesting system because it has relatively strong Fe II and Mg I absorption (see RTN06). These systems generally tend to have higher H i column densities (, RTN06), and therefore the identification of the galaxy causing this unusual absorption-line pattern might be illuminating. A complete optical and infrared set of images is available for this field. The images are shown in Figures 9 and 10. PSF subtractions were carried out on the optical data and no objects were detected within the subtracted region. The quasar PSF could not be subtracted on the infrared images as there were no suitable PSF stars in the field. Photometric measurements for the eight objects detected in this field are given in Tables 5 and 6.
Objects 1, 2, and 5 are in the SDSS database as having , , and respectively. Stickel et al. (1993) obtained a spectrum of Object 1, and determined it to be at redshift = 0.645. It is therefore ruled out as the absorber candidate. The best-fit stellar population synthesis models to our photometry for Objects 2, 4, and 8 are shown in Figure 11, and the best-fit template parameters are listed in Table 7. We derive a photometric redshift for Object 2, which is inconsistent with that listed in the SDSS database. The addition of infrared photometric measurements provides stronger constraints on the fit, making our redshift determination more reliable than that reported in the SDSS database. Object 3 is only detected in the infrared, and its = 0.29 and = 0.81 colours are not consistent with it being at the absorption redshift. We note that Object 3 has 7.6, which makes it an “extremely red object” (ERO). Object 4 is classified as a star in the SDSS database, however, it is extended on our images. The photometric redshift that we derive for Object 4 is consistent with the absorption redshift, and it is identified as the candidate absorber in this field. A stellar population synthesis model could not be fit to the photometry of Object 5 probably because of its low surface brightness and low detection significance in all our images. Its photometry is therefore highly uncertain.
In summary, Object 4 is selected as the absorbing galaxy in this field since its photometric redshift, , matches the absorption redshift, and the galaxies with smaller impact parameters are ruled out as candidate absorbers. This identification is assigned CL = 1.
|QSO||+0.0||0.0||0.0||15.71 0.001||74.8 (2007)||15.50 0.001||138.8 (2187)||15.90 0.001||225.1 (1231)||15.20 0.001||37.7 (4057)|
|1||5.2||+5.1||7.3||23.04 0.06||2.6 (67)||23.80 0.07||2.8 (51)||23.89 0.06||2.6 (67)||19.81 0.01||3.5 (633)|
|2||+7.7||3.2||8.4||23.21 0.07||2.3 (66)||24.46 0.09||2.8 (28)||24.06 0.07||2.3 (66)||21.62 0.03||3.3 (125)|
|4||1.4||+12.6||12.7||24.17 0.08||2.9 (36)||21.47 0.03||3.4 (140)|
|5||+10.8||10.1||14.8||23.57 0.08||2.5 (43)||23.16 0.05||3.0 (86)||24.41 0.08||2.5 (43)||22.93 0.07||2.4 (52)|
|6||+15.6||+0.2||15.6||25.33 0.18||2.1 (10)||24.88 0.11||4.5 (12)||26.17 0.18||2.1 (10)|
|7||+16.4||1.5||16.4||25.14 0.17||2.1 (12)||25.99 0.17||2.1 (12)|
|8||1.0||+17.3||17.3||24.73 0.15||3.4 (11)||23.72 0.06||2.4 (66)||25.58 0.15||3.4 (11)||24.39 0.14||2.3 (14)|
|QSO||+0.0||0.0||0.0||14.04 0.13||116.3 (578)||13.58 0.16||116.3 (433)||13.35 0.13||142.8 (535)|
|1||5.2||+5.1||7.3||19.40 0.13||4.3 (114)||18.99 0.16||4.1 (83)||18.66 0.13||5.6 (102)|
|2||+7.7||3.2||8.4||20.15 0.14||3.6 (67)||19.64 0.17||3.7 (51)||19.24 0.13||6.4 (53)|
|3||+0.1||9.0||9.0||21.31 0.15||3.3 (25)||21.02 0.19||2.6 (20)||20.21 0.14||4.9 (28)|
|4||1.4||+12.6||12.7||20.77 0.14||3.4 (40)||20.27 0.17||3.4 (31)||19.98 0.14||4.3 (40)|
|5||+10.8||10.1||14.8||21.88 0.18||2.1 (24)||22.17 0.21||2.5 (9)|
|Galaxy||Stellar Population Synthesis Model Parameters|
Example 3: the 1109+0051 field
This field is not as straightforward as the previous ones. The sightline towards the quasar 1109+0051 (SDSS J110936.35+005111.3) contains two subDLA systems, one at = 0.4181 with a column density of and the other at = 0.5520 with a column density of (RTN06). Images of this field were obtained in g’, r’, J, H, and K, from which six objects are detected (Figures 12 and 13). The quasar PSF subtraction revealed no object within the subtracted region. The optical and infrared photometry are given in Tables 8 and 9, respectively.
Object 1 has an impact parameter , which at the two absorber redshifts corresponds to 7 kpc and 8 kpc, respectively. It is detected in the , and -bands. Since it overlaps with the quasar PSF, its photometry is uncertain. We consider it to qualify as a candidate absorber due to its proximity to the sightline. The best-fit stellar population synthesis models to our photometry for Objects 2, 3, 4, and 5 are shown in Figure 14, and the model parameters are tabulated in Table 10. The best-fit stellar population synthesis model to the photometry of Object 2 is only marginally consistent with (within of) the lower absorption redshift system, = 0.4181. The photometry of Object 3 results in a photometric redshift of , which is consistent with the absorption system at = 0.5520. However, as can be seen from Table 10, the photometric redshift we derive for Object 4, while inconsistent with the SDSS-derived photometric redshift of , is consistent with both absorption redshifts. In addition, the redshift derived for Object 5 is consistent with the absorption system at = 0.4181 (but inconsistent with the SDSS redshift of ). Therefore, Objects 1, 3, 4, and 5 are all potential absorber candidates for the absorption systems in this field, while Object 2 is marginally consistent at the lower redshift. Given this ambiguity, we use the proximity criterion as the deciding factor, and select Object 1 as the = 0.4181 candidate and Object 3 as the = 0.5520 candidate, both with confidence level CL = 2.
|QSO||0.0||0.0||0.0||18.45 0.03||29.5 (422)||18.45 0.05||9.1 (263)|
|1||0.6||1.1||1.3||23.79 0.33||2.6 (26)|
|2||3.4||7.4||8.1||24.58 0.64||2.6 (12)||22.79 0.42||2.7 (14)|
|3||+8.4||5.4||10.0||24.29 0.56||2.6 (16)||23.54 0.67||2.3 (8)|
|4||+4.6||9.9||10.9||23.15 0.24||3.8 (34)||22.05 0.27||3.8 (20)|
|5||+13.4||0.7||13.4||21.44 0.07||4.8 (142)||20.52 0.10||3.0 (111)|
|6||11.7||10.6||15.8||22.85 0.42||2.2 (16)|
|QSO||0.0||0.0||0.0||18.06 0.01||12.8 (151)||18.41 0.01||9.1 (113)||18.03 0.01||10.8 (119)|
|1||0.6||1.1||1.3||22.24 0.13||2.6 (16)||22.36 0.17||2.1 (13)|
|2||3.4||7.4 0.||8.1||21.07 0.07||2.6 (47)||21.10 0.08||2.5 (34)||20.18 0.05||3.2 (55)|
|3||+8.4||5.4||10.0||23.98 0.29||1.7 (5)||22.63 0.19||2.1 (10)||22.88 0.23||2.4 (6)|
|4||+4.6||9.9||10.9||22.67 0.16||2.0 (14)||22.86 0.21||2.5 (6)|
|5||+13.4||0.7||13.4||20.35 0.04||3.5 (67)||20.24 0.05||3.2 (59)||20.17 0.05||3.3 (55)|
|Galaxy||Stellar Population Synthesis Model Parameters|
Example 4: the 1715+5747 field
This is a case where no galaxy is identified as the absorber. The sightline towards the quasar 1715+5747 (SDSS J171539.86+574722.2) contains a subDLA system at = 0.5579 with a column density of (RTN06). A complete optical and infrared dataset was obtained for this field. Figures 15 and 16 show that only three objects are detected within 100 kpc of the quasar at . PSF subtractions were carried out on the optical images and no objects were detected within the subtracted region. The quasar PSF could not be subtracted on the infrared images as there were no suitable PSF stars in the field. Object 1 is located 3.5 from the quasar sightline which corresponds to 22.3 kpc at the absorption redshift. Objects 2 and 3 have and , respectively, according to the SDSS database.
The best-fit stellar population synthesis model to our photometry for Objects 1, 2, and 3 are shown in Figure 17, and the model parameters are listed in Table 13. The photometric redshift we derive for Object 1 does not match the absorption redshift. While the SDSS photometric redshift for Object 2 is consistent with the redshift of the absorption system and our optical photometric measurements agree well with those measured by the SDSS, the addition of our IR data results in a very different photometric redshift for Object 2. Again, as was the case for Object 2 in the 0735+178 field, the addition of IR data was crucial for the determination of the galaxy’s redshift. With regards to Object 3, we derive a photometric redshift that is consistent with the one obtained by the SDSS.
Thus, none of the objects detected in this field have photometric redshifts consistent with the absorption redshift. Based on their proximity to the quasar sightline, one might expect either Object 1 or 2 to be the absorbing galaxy. However, until spectroscopic data or better photometry are available that might prove our results to be incorrect, we consider our data on this field to be inconclusive, i.e., we do not have an absorbing galaxy identification. It may be one of the cases where the absorbing galaxy is (5.2 kpc) from the quasar sightline, or at a larger impact parameter and fainter than the brightness limit of our K-band data, = 0.09.
|QSO||0.0||0.0||0.0||19.69 0.15||21.9 (496)||18.21 0.04||40.5 (758)||18.44 0.04||18.4 (770)||18.55 0.05||20.5 (424)|
|1||1.4||3.2||3.5||24.87 0.41||2.1 (34)||24.76 0.53||2.0 (21)||21.56 0.08||3.5 (160)|
|2||4.3||+0.9||4.4||23.99 0.27||2.7 (78)||22.07 0.07||4.1 (224)||21.07 0.05||4.4 (282)||20.74 0.06||5.0 (230)|
|3||+3.2||+11.1||11.5||22.46 0.19||3.3 (266)||20.49 0.05||8.4 (456)||19.51 0.04||8.4 (630)||19.20 0.05||8.8 (539)|
|QSO||0.0||0.0||0.0||18.07 0.01||11.8 (82)||17.82 0.01||12.2 (103)||17.28 0.01||16.0 (139)|
|1||1.4||3.2||3.5||20.54 0.08||2.8 (35)||20.21 0.06||3.4 (41)||19.82 0.04||3.2 (68)|
|2||4.3||+0.9||4.4||20.13 0.06||3.2 (46)||19.69 0.04||3.7 (61)||19.48 0.03||4.0 (74)|
|3||+3.2||+11.1||11.5||18.57 0.03||3.5 (174)||18.17 0.02||4.6 (196)||17.92 0.01||5.0 (249)|
|Galaxy||Stellar Population Synthesis Model Parameters|
4 Galaxy Properties
4.1 Earlier work
Absorber galaxies that had been previously identified are presented in Table 14. The first results from our DLA imaging programme that were presented in Turnshek et al. (2001) and Rao et al. (2003) are included in this table. Additionally, our current sample (Table 1) has five fields that were studied by other investigators, however, our new images did not alter the earlier conclusions. These five are also included in Table 14 with previous studies referenced.
Mg ii absorbers from RTN06 are tabulated in the first section of Table 14, and those not in RTN06 are included in the second section of the table. Column 1 is the quasar designation, column 2 gives the quasar emission redshift, columns 3 and 4, the Mg ii absorption redshift and rest equivalent width, and column 5, the H i column density of the absorber. Column 6 gives the impact parameter of the identified galaxy in kpc, and columns 7 and 8 give the galaxy’s AB magnitude and absolute luminosity with respect to (see §4.2). The relevant filter is noted in parentheses. The reference for the galaxy’s parameters is given in column 9 and the method by which it was identified is given in column 10. “Specz” indicates that a spectroscopic redshift was used to identify the galaxy, “Photoz” indicates that the galaxy’s photometric redshift matched the absorption redshift, and “Prox” indicates that the closest galaxy to the quasar sightline was chosen as the absorbing galaxy. Column 11 indicates the confidence level, CL, assigned to the identification (see §3). “Photoz” and “Prox” identifications are assigned CL = 1 and CL = 2, respectively. Specz is assigned CL = 1, except for the galaxy in the 1622+239 field. Steidel et al. (1997) obtained a spectroscopic redshift that matches for the galaxy at kpc in this field, but commented that a galaxy this faint and this far away could not be the DLA absorber. More recently, Kacprzak et al. (2007) have identified this galaxy as the absorber in their work, and we have adopted this new interpretation as well. However, we have assigned this galaxy an identification confidence level of CL = 2, because we feel that we cannot be as confident about the validity of this identification as we are about the rest of the spectroscopically-identified galaxy candidates (also see §6).
|MgII Systems from RTN06:|
4.2 Current work
Table LABEL:IDsummary provides details on DLA candidate galaxies in the 55 quasar fields that appear here for the first time. The first five columns are as described above for Table 14. We note here that column density values less than cm were obtained by fitting Voigt profiles with the same Doppler broadening parameter as the stronger subDLA and DLA lines (Rao & Turnshek 2000; RTN06). The column densities should therefore be considered approximate, but less than cm. Nevertheless, these are legitimate Mg ii absorption systems for which absorbing galaxies have been identified. Column 6 gives the object in each field that was identified as the absorbing galaxy candidate. As indicated in §3, the numbering is in order of increasing distance from the quasar sightline. Columns 7 and 8 are the galaxy’s impact parameter in arcsec and kpc, respectively, and column 9 gives the galaxy’s photometric redshift and associated error, which were determined if the field was observed through four or more filters.
Columns 10 and 11 give the galaxy’s AB magnitude and absolute luminosity with respect to . -band AB magnitudes and luminosities are provided unless the object was not observed (or detected) in , in which case the non- filter is noted. Our magnitude errors are typically 10 to 20%. See, for example, the photometry tables of individual fields in §3.
Apparent magnitudes are provided, unless otherwise noted. The
symbol “:” indicates that the magnitude is uncertain because the
galaxy overlaps with the quasar PSF.
Observations not photometric.
“BestTemplate” fit calculated by fixing the redshift of the stellar population template at the absorption redshift, to illustrate that a galaxy template consistent with the measured photometry exists.
Measurements of have been changed from RTN06 values to reflect the more recent measurements of Quider et al. (2011).
SDSS photometric redshift, -converted SDSS magnitudes are provided.
Measurements of have been changed from RTN06 values to reflect the more recent measurements of Mathes et al. (in preparation).
SDSS magnitudes provided by S. Zibetti (private communication).
The following values were used to determine values for the galaxies listed in Tables 14 and LABEL:IDsummary in the filters indicated:
(, Dahlen et al. 2005)
(, Dahlen et al. 2005)
(, Gabasch et al. 2004)
(, Gabasch et al. 2004)
(Dahlen et al. 2005)
(, Wolf et al. 2003)
(, Ilbert et al. 2005)
(, Ilbert et al. 2005)
(, Dahlen et al. 2005)
(, Dahlen et al. 2005)
(Jones et al. 2006; assuming no evolution between and 0.5)
(Cirasuolo et al. 2006)
-corrections for galaxies whose redshifts were obtained using template fits to the photometry were determined from the template fits themselves. For the rest, an Sb-type -correction in the observed filter at the redshift of the absorber was assumed.
Columns 12 and 13 give the method by which the DLA galaxy candidate was identified, and CL, the confidence level of this identification (§3). Of the 66 absorbers in Table LABEL:IDsummary, 17 have photometric redshifts that match the absorption redshift, and are assigned CL . Thirty-seven identifications were made either with colours that were consistent with the galaxy being at the absorption redshift, the proximity criterion, or photometric-redshift matches that were only marginally consistent with the absorption redshift. These are labeled as having CL = 2 or 3, with 2 being the more confident identification. No galaxy identification was possible for 12 absorbers. Examples of some of these were given in §3.
Figure 18 shows that the CL = 1 and CL = 2 or 3 samples have very similar impact parameter and luminosity distributions. Here, we have included galaxies from Tables 14 and LABEL:IDsummary. In Figure 18, impact parameter, , is plotted versus log H i column density for the 80 galaxies that have and measurements. Galaxy luminosity is represented by the size of the symbol (see caption). We will discuss the - plane in more detail later, however, this plot clearly shows that the CL values do not cluster with either parameter, or with luminosity. In addition, Kolmogorov-Smirnov (KS) tests show that the two CL samples are drawn from the same parent population: the KS test probabilities are 0.08 for the two luminosity distributions, 0.68 for the two impact parameter distributions, and 0.94 for the two H i column density distributions. Therefore, this is evidence that we are statistically selecting similar candidate galaxies in all of these samples. Hereafter, we will no longer separate the sample by CL value, and will explore the properties of all candidate galaxies irrespective of their indentification method.
Our galaxy sample is essentially defined by five parameters:
absorption redshift, , Mg ii 2796 rest equivalent
width, , H i column density, , galaxy
impact parameter, , and galaxy luminosity, which we express as a
fraction of , . The three galaxy identifications from
Table 15 that have measured values but no measurements are
not included. The sample we analyse includes 80 absorption systems and
their identified galaxies. Figure 19 shows the
distributions of these properties at a glance. Open circles are
systems with and solid circles are DLAs. In
Table LABEL:Spearman, we provide results from the Spearman rank
correlation test in order to quantify possible correlations among the
In addition, Table 17 lists KS test probabilities that two
given samples are drawn from the same parent population. In the first
set, , , in kpc, and are
compared for three different samples of H i column density: the DLA
() and subDLA ()
samples, and for the subDLA and Lyman limit system (LLS, ) samples. There are 27 galaxies in the DLA sample, 30 in the
subDLA sample, and 23 in the LLS sample. Next, the sample is split by
the median value of
|Parameter||DLA vs. SubDLA
Å vs. Å