Probing the IGM-Galaxy Connection at z < 0.5 I: A Galaxy Survey in QSO Fields and a Galaxy-Absorber Cross-Correlation Study
aa r X i v : . [ a s t r o - ph . C O ] J un A CCEPTED FOR P UBLICATION IN THE A STROPHYSICAL J OURNAL
Preprint typeset using L A TEX style emulateapj v. 4/12/04
PROBING THE IGM-GALAXY CONNECTION AT Z < , H SIAO -W EN C HEN AND J OHN
S. M
ULCHAEY Accepted for Publication in the Astrophysical Journal
ABSTRACTWe present an imaging and spectroscopic survey of galaxies in fields around QSOs HE 0226 - - + α absorbers and 13 O VI absorbers along the three sightlines. We obtain robust redshifts for1104 galaxies of rest-frame absolute magnitude M R - h . -
16 and at projected physical distances ρ . h - Mpc from the QSOs. HST/WFPC2 images of the fields around PKS 0405 -
123 and PG 1216 +
069 are availablefor studying the optical morphologies of absorbing galaxies. Combining the absorber and galaxy data, weperform a cross-correlation study to understand the physical origin of Ly α and O VI absorbers and to constrainthe properties of extended gas around galaxies. The results of our study are (1) both strong Ly α absorbersof log N (H I) ≥
14 and O VI absorbers exhibit a comparable clustering amplitude as emission-line dominatedgalaxies and a factor of ≈ r p < h - Mpc; (2) weak Ly α absorbers of log N (H I) < . r p ≤ h - kpc has a correspondingO VI absorber to a sensitive upper limit of W (1031) . .
03 Å, while the covering fraction of O VI absorbing gasaround emission-line dominated galaxies is found to be κ ≈ W (1031) >
70 mÅ.
Subject headings: cosmology: observations—intergalactic medium—quasars: absorption lines
1. INTRODUCTION
Traditional galaxy surveys trace large-scale structures visi-ble through stellar light or radio emission of cold neutral gas.Stars and known gaseous components account for roughly12% of all baryons in the local universe and the rest arethought to reside in ionized gaseous halos around galaxiesor in intergalactic space (e.g. Fukugita 2004). In principle,the diffuse halo gas and intergalactic medium (IGM) can beprobed by the forest of absorption line systems observed inthe spectra of background QSOs. Studies of QSO absorption-line systems are therefore expected to provide important con-straints for theoretical models that characterize the growth oflarge scale structures.Indeed, the Ly α forest is found to account for & z = 2 - α absorbers of neutral hy-drogen column density N (H I) ≤ . cm - may contain be-tween 20 -
30% of the total baryons based on a simple as-sumption for the cloud geometry. Various numerical sim- Based in part on observations made with the NASA/ESA Hubble SpaceTelescope, obtained at the Space Telescope Science Institute, which is op-erated by the Association of Universities for Research in Astronomy, Inc.,under NASA contract NAS 5-26555. Observations reported here were obtained in part at the Magellan tele-scopes, a collaboration between the Observatories of the Carnegie Institu-tion of Washington, University of Arizona, Harvard University, University ofMichigan, and Massachusetts Institute of Technology. Dept. of Astronomy & Astrophysics and Kavli Institute for Cos-mological Physics, University of Chicago, Chicago, IL, 60637, U.S.A. [email protected] The Observatories of the Carnegie Institution of Washington 813 SantaBarbara Street, Pasadena, CA 91101, U.S.A. [email protected] ulations have further suggested that approximately 40% ofthe total baryons reside in diffuse intergalactic gas of tem-perature T < - K (e.g. Davé et al. 2001; Cen & Ostriker2006). This temperature range makes high-ionization tran-sitions, such as O VI λλ λλ ρ = 100 h - kpc from a background QSOhave associated absorbers produced by C IV λλ λλ ρ = 180 h - kpc have as-sociated Ly α absorbers of N (H I) ≥ cm - (Chen et al.1998, 2001a), not all Ly α absorbers have a galaxy foundwithin 1 h - Mpc physical distance to a luminosity limit of0 . L ∗ (Morris et al. 1993; Tripp et al. 1998; Stocke et al.2006). It is not clear whether these absorbers are associatedwith fainter galaxies or not related to galaxies at all. In ad-dition, the origin of O VI absorbers in terms of their galac- Chen & Mulchaeytic environment is also poorly quantified. While Stocke et al.(2006) found 95% of O VI absorbers at z < . ρ . h - kpc from an L ∗ galaxy, only three of the six O VIabsorbers studied in Prochaska et al. (2006) have an L ∗ galaxyfound at ρ < h - Mpc.To understand the origin of QSO absorption-line systems,Chen et al. (2005) presented a pilot study of the galaxy–Ly α absorber cross-correlation function based on the Ly α absorbers and galaxies identified along a single sightline to-ward PKS0405 - ξ ga including only emission-line domi-nated galaxies and strong Ly α absorbers of log N (H I) ≥ ξ gg on co-moving, projection distance scales r p ≤ h - Mpc, while there remained a lack of cross-correlationsignal when using only absorption-line dominated galaxies.Absorption-line dominated galaxies have red colers and arepresumably evolved early-type galaxies, while emission-linedominated galaxies are blue and presumably younger star-forming systems. Early-type galaxies are found to clustermore strongly than younger, star-formig galaxies (e.g. Madg-wick et al. 2003; Zehavi et al. 2005), and are therefore ex-pected to reside in regions of higher matter overdensity. Thecomparable correlation amplitudes of emission-line galaxiesand strong Ly α absorbers therefore suggested that strong ab-sorbers of log N (H I) ≥
14 reside in the same overdensity re-gions as emission-line dominated galaxies. It appeared thatthese absorbers do not trace regions where more massivegalaxies with dominant absorption spectral features reside.This result of Chen et al. (2005) provided the first quan-titative constraint on the origin of Ly α absorbers in termsof the significance of the underlying matter density fluctua-tions around the regions where they reside. It also offereda physical explanation for the on-average weaker clusteringamplitude of Ly α absorbers relative to the clustering ampli-tude of the luminous galaxy population reported earlier (e.g.Morris et al. 1993). Wilman et al. (2007) performed a simi-lar analysis based on a galaxy and Ly α absorber pair sampleestablished over 16 QSO sightlines. While these authors con-cluded that a spectral-type dependent galaxy and Ly α cross-correlation function was not confirmed by their analysis, thetwo-dimensional cross-correlation function presented in Fig-ure 4 of Wilman et al. clearly displays a strong signal betweenLy α absorbers and emission-line galaxies that is absent in thecross-correlation function measured using only absorption-line galaxies.To examine whether the initial results of Chen et al. (2005)obtained based on a single field are representative of the gen-eral Ly α absorber population, we have been conducting adeep, wide-area survey of galaxies in fields around QSOs at z = 0 . - .
6. The QSO fields are selected to have ultravi-olet echelle spectra available from the Far Ultraviolet Spec-troscopic Explorer (FUSE) and the Space Telescope Imag-ing Spectrograph (STIS) on board the Hubble Space Tele-scope (HST). The high-resolution UV spectra are necessaryfor finding intervening hydrogen Ly α and O VI absorbers toform a statistically representative sample (see e.g. Thom & Chen 2008a; Tripp et al. 2008), as well as identifying theirassociated metal-line transitions for constraining the ioniza-tion state of the gas. The galaxy sample from our survey pro-gram therefore also allows us to expand upon the initial Ly α absorber study to understanding the galactic environment ofionized gas traced by the O VI absorbers. The primary ob-jectives of our galaxy survey program are (i) to examine thephysical origin of Ly α and O VI absorbers based on their re-spective clustering amplitudes, and (ii) to constrain the prop-erties of extended gas around galaxies.This paper is organized as follows. In Section 2, we de-scribe the design of our galaxy survey program. In Sec-tion 3, we describe our observing program that includes bothimaging observations, the selection of candidate galaxies, andmulti-slit spectroscopic observations of three QSO fields. InSection 4, we summarize the results of the spectroscopic sur-vey and examine the survey completeness. In Section 5, wesummarize the properties of known absorption-line systemsalong individual QSO sightlines and present a new Ly α ab-sorber catalog for the sightline toward PG 1216 + Λ CDM cosmology, Ω M = 0 . Ω Λ = 0 .
7, with a dimensionless Hubble constant h = H / (100 km s - Mpc - ) throughout the paper.
2. GOALS AND DESIGN OF THE GALAXY SURVEY
The primary objectives of our galaxy survey program are(i) to examine the physical origin of Ly α and O VI absorbersbased on their clustering amplitudes, and (ii) to constrain theproperties of extended gas around galaxies. To reach thegoals, we have been conducting a wide-area survey of galax-ies in fields around QSOs at z = 0 . - .
6, for which echellespectra at ultraviolet wavelengths are available from FUSEand HST/STIS. The high-resolution UV spectra are necessaryfor finding intervening hydrogen Ly α and O VI absorbers toform a statistically representative sample (see e.g. Thom &Chen 2008a; Tripp et al. 2008), as well as identifying their as-sociated metal-line transitions for constraining the ionizationstate of the gas. The clustering amplitudes of Ly α and O VIabsorbers are then determined based on their cross-correlationsignals with the galaxies identified in the wide-field redshiftsurvey.To characterize the cross-correlation function of galaxiesand Ly α /O VI absorbers, we aim to establish a statisticallyrepresentative sample of L ∗ galaxies on scales of ≈ - h - co-moving Mpc from the QSO lines of sight over a redshiftrange of z = 0 . - .
5. These luminous galaxies are thought toreside at the peaks of the underlying dark matter density fluc-tuations with a small fraction ( . Λ CDM paradigm, the relative clus-tering amplitudes of QSO absorption-line systems with re-spect to these galaxies provide quantitative measures of theoverdensities of the regions where the absorbers reside (Mo& White 2002; Tinker et al. 2005).To examine the properties of extended gas around galax-ies, we aim to establish a complete sample of sub- L ∗ galax-ies at z = 0 . - . r p ≈ h - kpc of the QSO lines of sight. The galaxy andabsorber pair sample allows us to study the incidence and ex-orrelation of Galaxies and QSO Absorption-line Systems 3tent of ionized gas around galaxies of a range of luminosityand stellar content.The short camera in the IMACS multi-object imaging spec-trograph (Dressler et al. 2006) on the Magellan Baade tele-scope contains eight CCDs and has a plate scale of 0 . ′′ .It observes a sky area of ≈ ′ radius, corresponding to aco-moving distance of 3 . h - Mpc at z = 0 .
3. In a singlesetup, IMACS provides the necessary field coverage to effi-ciently carry out our galaxy survey program. To reach thescientific objectives outlined above, we have selected threeQSOs, HE 0226 - z QSO = 0 . -
123 ( z QSO =0 . +
069 ( z QSO = 0 . >
80% completeness for galaxies brighter than R = 22 and atan angular radius less than ∆ θ = 2 ′ from the QSOs and (2) &
50% completeness for galaxies brighter than R = 20 and at ∆ θ > ′ . At z = 0 . R = 22 corresponds to ≈ . L ∗ galaxiesand ∆ θ = 2 ′ corresponds to ρ = 280 h - kpc. At z = 0 . R = 22corresponds to ≈ . L ∗ galaxies and ∆ θ = 2 ′ corresponds to ρ = 510 h - kpc. We note that the field around PKS 0405 - R = 20, which was only sensitive to roughly L ∗ galaxies at z = 0 .
5. Our program has been designed to expand upon theearlier spectroscopic survey for measuring redshifts of faintergalaxies along the QSO line of sight.
3. OBSERVING PROGRAM
The observing program of the galaxy survey consists ofthree phases: (1) imaging observations of the QSO fields toidentify candidate galaxies, (2) object selection for follow-upfaint galaxy spectroscopy, and (3) multi-slit spectroscopic ob-servations to measure the redshifts of the selected galaxies. Inthis section, we describe each of the three phases.3.1.
Imaging Observations and Data Reduction
Imaging observations of the fields around HE 0226 - +
069 were necessary in order to identify galax-ies for follow-up spectroscopy . Faint galaxies in the fieldaround PKS 0405 -
123 have been published in Prochaskaet al. (2006), which are complete to R = 22 .
5. We have there-fore selected our spectroscopic targets of this field from thisgalaxy catalog.Optical images of the field around HE 0226 - B , R , and I filters on the Magellan Baade Tele-scope. Optical images of the field around PG 1216 +
069 wereobtained on 2005 March 19, using the IMACS short cam-era and the B , R , and I filters. To increase the efficiency inthe spectroscopic follow-up, the multi-bandpass B , R , and I We note that the field around PG 1216 +
069 is covered by the Sloan Digi-tal Sky Survey (SDSS; York et al. 2000), which has a targeted imaging depthof r ′ = 22 . h FWHM i = 1 . ′′ , the available SDSS imaging data become significantly in-complete for detecting distant faint galaxies near the magnitude limit and forresolving objects near the QSO. Additional imaging data are therefore neces-sary for the purpose of our study. photometric measurements were necessary for selecting can-didate galaxies at z < .
5, excluding contaminating red starsand possible luminous high-redshift galaxies. The observa-tions were carried out in a series of four to six exposures ofbetween 300 and 500 s each, and were dithered by between20 and 30 arcsec in space to cover the chip gaps between in-dividual CCDs. All the imaging data were obtained underphotometric conditions. We also observed standard star fieldsselected from Landolt (1992) through each of the B , R , and I filters on each night in order to calibrate the photometry inour targeted fields. Flat-field images were taken both at a flatscreen at the secondary mirror of the Magellan Baade tele-scope and at a blank sky during evening twilight. A journalof the IMACS imaging observations is presented in Table 1,which lists in columns (1) through (6) the object field, instru-ment, filter, total exposure time, mean FWHM, and Date ofobservations, respectively.The reduction and processing of IMACS imaging data werecomplicated because of the multi-CCD format and becauseof a substantial geometric distortion across the field of theIMACS short camera. Before stacking individual exposures,we first formed a single, geometrically-dewarped frame of ev-ery exposure according to the following steps. First, we sub-tracted the readout bias of individual CCDs using the overscanregions of the chips. Next, the bias subtracted frames wereflattened using a super sky flat formed from median-filteringunregistered science exposures. We found that in the B and R images a super sky flat works more effectively than eitherdome flats or twilight flats in removing the gain variationsbetween individual pixels. For images obtained through the I filter, however, the fringes in individual science exposuresmade it impossible to obtain an accurate flat-field image fromthe data. We therefore flattened the I -band exposures by firstcorrecting the pixel-to-pixel variation using a median domeflat. The remaining fringes were then removed by subtractinga median I -band sky image formed from individual flattenedscience exposures.Next we registered individual CCDs of each exposure us-ing known stars found in the USNO A2.0 catalog. We formeda geometrically corrected mosaic frame of eight chips usingthe IRAF MSCRED package. The resulting mosaic framecontained calibrated astrometry with a typical r.m.s. scatterin stellar positions of 0 . ′′ . Finally, we combined individualmosaic frames to form a single stacked image of the QSOfield for each of the BRI bandpasses. The final stacked im-ages covered a contiguous sky area of ≈ ×
28 arcmin and have mean point spread functions (PSFs) in the range ofFHWM = 0 . ′′ - . ′′ in the center of the IMACS field. Photo-metric calibrations were obtained using Landolt standard starsobserved on the same night. The combined mosaic imagesreached 5 σ limiting Vega magnitudes of B = 24 . R = 24 . I = 23 .
5. False color images of the central 4 ′ × ′ re-gions around HE 0226 - +
069 are presentedin Figures 1 and 2, respectively.Additional optical images of PKS 0405 -
123 andPG1216 +
069 obtained with HST using the Wide Fieldand Planetary Camera2 (WFPC2) with the F702W filterwere accessed from the Hubble Space Telescope (HST) dataarchive. Individual exposures of each field were reducedusing standard pipeline techniques, registered to a commonorigin, filtered for cosmic-ray events, and stacked to form afinal combined image. These high spatial resolution imagesallowed us to resolve faint galaxies close to the QSO line ofsight and to study their morphology in detail (see Figures 8 Chen & Mulchaey
TABLE 1J
OURNAL OF I MAGING O BSERVATIONS
Exposure FWHMField Instrument Filter Time (s) (arcsec) Date(1) (2) (3) (4) (5) (6)HE 0226 - z QSO = 0 . B R I -
123 ( z QSO = 0 . +
069 ( z QSO = 0 . B R I IG . 1.— A combined IMACS R -band image of the center 4 ′ × ′ field around HE 0226 - z QSO = 0 . & 9 below). A journal of the HST/WFPC2 observations ispresented in Table 1 as well.3.2. Selection of Candidate Galaxies for SpectroscopicFollow-up
To select galaxies in the fields around HE 0226 - +
069 for spectroscopic follow-up, we first identifiedobjects in the stacked R -band mosaic image using the SEx-tractor program (Bertin & Arnout 1996). Next, we measuredobject magnitudes in the B , R , and I images using the isopho-orrelation of Galaxies and QSO Absorption-line Systems 5 F IG . 2.— A combined IMACS R -band image of the center 4 ′ × ′ field around PG 1216 +
069 ( z QSO = 0 . ′′ northeast of the QSO makes it challenging to accurately identify faint galaxies near the QSO sightline. tal apertures of individual objects determined by SExtractor.This procedure yielded 948 and 1619 objects of R = 16 - ∆ θ = 11 ′ angular radius from HE 0226 - + +
069 is known to passthrough the outskirts of the Virgo cluster at z = 0 .
004 (and aprojected physical distance ρ ≈ . h - Mpc). As discussedbelow, the results of our spectroscopic survey have also un-covered multiple large-scale overdensities along the sightlinethat contribute to the high surface density of galaxies in thisfield.Separating stars and galaxies was challenging in these im-ages, because of non-negligible PSF distortions toward theedge of the IMACS field. To optimize the efficiency of thespectroscopic observations, we applied the observed B - R versus R - I colors to exclude likely contaminations fromfaint red stars and high-redshift galaxies. Figure 3 showsthe B - R vs. R - I color distribution of the objects found inthe field around HE 0226 - z = 0 (lower left) to z = 1 (upper right) in steps of ∆ z = 0 . z . . R = 16 -
22 and B - R + . > . R - I ) within ∆ θ = 11 ′ angular radius fromHE 0226 - + L ∗ galaxiesfor tracing the large-scale matter overdensities along the linesof sight toward the QSOs and (2) to identify a highly com-plete sample of sub- L ∗ galaxies within projected co-movingdistance r p ≈ h - kpc from the QSO lines of sight forstudying extended gas around individual galaxies and galaxygroups. In designing the multi-slit masks, we therefore aimedto include nearly all galaxies within ∆ θ = 2 ′ of the back-ground QSO. Because of spectral collisions between closelylocated objects, this goal dictates the number of masks re-quired to reach a high survey completeness. We were able Chen & Mulchaey F IG . 3.— The B - R vs. R - I color distribution of the objects found in thefield around HE 0226 - z = 0 (lower left) to z = 1 (upper right) in stepsof ∆ z = 0 . z . . to reach the goal in five IMACS masks and included 693and 750 objects in the multi-slit spectroscopic observationsof HE 0226 - + .For the field around PKS 0405 - ∆ θ = 11 ′ of the QSO and brighter than R =22 from the photometric catalog of Prochaska et al. (2006).These galaxies had not been observed in the previous sur-vey. We aimed to complete the redshift survey of galaxiesto R ≤
22 in this QSO field with IMACS observations. Wedesigned five IMACS masks to observe 450 objects in thisfield.3.3.
Spectroscopic Observations and Data Reduction
Multi-slit spectroscopic observations of objects selectedfrom the imaging program described in § 3.2 were carriedout using IMACS and the short camera during three dif-ferent runs in November 2004, May 2006, and April 2007.We were able to complete the observations of five masksfor the field around HE 0226 - - + . ′′ in width. Typical slit lengths were five arcsec-onds for compact sources. For extended objects, we adoptedthe Kron radius estimated in SExtractor and expanded by afactor of three to allow accurate sky subtraction. We usedthe 200 l/mm grism that offers a spectral coverage of λ =5000 - ≈ Some objects were duplicated in different masks. hour for wavelength and flat-field calibrations. A journal ofthe IMACS spectroscopic observations is presented in Table2, which lists in columns (1) through (7) the object field, in-strument set-up, spectral resolution (FWHM), mask number,number of objects observed per mask, total exposure time,and Date of observations, respectively.Additional spectroscopic observations of 27 objects at ∆ θ < ′ from HE 0226 - -
123 were attempted in October 2007 and Decem-ber 2008, using the Low Dispersion Survey Spectrograph 3(LDSS3) on the Magellan Clay telescope. We included galax-ies fainter than R = 22 and clearly visible near the QSO sight-lines in the LDSS3 observations, to further improve the sur-vey depth in the immediate vicinity of the QSOs. LDSS3 ob-serves a field of 8 . ′ diameter at a pixel scale of 0 . ′′ . Weused the volume phase holographic (VPH) blue and VPH all grisms that cover a spectral range of λ = 4000 - λ = 6000 - . ′′ × ′′ slitlets. Theadditional LDSS3 multi-slit spectroscopic observations allowus to reach 100% survey completeness of faint galaxies nearthe QSOs. A journal of the LDSS3 observations is includedin Table 2 as well.The multi-slit spectroscopic data were processed and re-duced using the Carnegie Observatories System for MultiOb-ject Spectroscopy (COSMOS) program that has been devel-oped and tested by A. Oemler, K. Clardy, D. Kelson, and G.Walth. The program adopts the optical model of the spec-trograph and makes initial guesses both for the slit locationsin the 2D spectral frame and for the wavelength calibrationsof individual spectra. These initial gueses are then furtherrefined using known spectral lines in the He-Ne-Ar frame ob-tained close in time with the science data. The rms scatter ofthe wavelength solution is typically a fraction of a pixel, i.e. .
4. RESULTS OF THE SPECTROSCOPIC SURVEY
Redshift Measurements and Spectral Classification
The redshifts of objects were determined independently byeach of us in two steps. First, we performed separate au-tomatic χ fitting routines that determine a best-fit redshiftbased on cross-correlating the flux-calibrated object spectrumwith an input model template. One of the cross-correlationroutines was written by D. Kelson. This routine adopts SDSSspectral templates for early-type (SDSS template > ,
000 SDSS galaxy spectra, us- orrelation of Galaxies and QSO Absorption-line Systems 7 TABLE 2J
OURNAL OF M ULTI - SLIT S PECTROSCOPIC O BSERVATIONS
Instrument/ FHWM ExposureField Disperser Å Mask No. of Slitlets Time (s) Date(1) (2) (3) (4) (5) (6) (7)HE 0226 - z QSO = 0 . -
123 ( z QSO = 0 . +
069 ( z QSO = 0 . IG . 4.— Relative significances between absorption ( C ) and emission( C ) components for 220 galaxies in the field around PKS 0405 - C is determined by the best-fit coefficient of the firsteigen spectrum. The emission coefficient C is determined by the best-fitcoefficient of the second eigen spectrum, corrected for the continuum slopeby subtracting off the best-fit coefficients of the remaining two eigen spectra. ing a principal component analysis (see e.g. Yip et al. 2004).The first two eigen spectra are characterized by predominantlyabsorption features and predominantly emission features, re-spectively. The last two eigen spectra offer additional modi-fications in the continuum slope. A linear combination of thefour eigen function formed a model template to be comparedwith a flux-calibrated object spectrum, and a χ routine wasperformed to determine the best-fit linear coefficients and red-shift for each object.Next, the best-fit redshifts returned from the cross- correlation routines were then visually inspected by each ofus to determine a final and robust redshift for every object.In some cases, the cross-correlation routines failed to iden-tify a correct redshift due to contaminating residuals of brightskylines or imperfect background sky subtraction. We wereable to recover the redshifts during this visual inspection pro-cess based on the presence of Ca II H&K, or the presenceof H α and [N II]. A comparison between redshifts deter-mined independently by the two of us shows a typical scat-ter of | ∆ z | ≈ . σ uncertainty of 70km s - .The procedure described above led to robust redshift mea-surements for 432 and 448 galaxies in the fields aroundHE 0226 - + - S / N in their spectra for arobust redshift measurement.Spectral classification was guided by the relative magni-tudes between the best-fit coefficients of the eigen spectradescribed above. These coefficients provide a quantitativeevaluation of the fractional contributions from different spec-tral components in a galaxy, such as absorption spectra dueto low-mass stars or emission spectra due to H II regions.The contribution due to absorption components may be de-termined by the best-fit coefficient of the first eigen spec-trum ( C ). The contribution due to emission components maybe determined by the best-fit coefficient of the second eigenspectrum, corrected for the continuum slope by subtractingoff the best-fit coefficients of the remaining two eigen spec-tra ( C ). Figure 4 shows that despite a large scatter, thereexists a linear distribution between emission C and absorp-tion C components for 224 galaxies found in the field aroundPKS 0405 - C > .
09 and C < . Survey Completeness and Galaxy Redshift Distribution
Chen & MulchaeyThe results of our spectroscopic survey program in threeQSO fields are summarized in Figure 5. For the field aroundHE 0226 - R = 23 and the closed circles indi-cate the ones with available spectroscopic redshifts. The QSOis at RA(J2000) = 37 . - . -
123 and PG 1216 + R = 22 and the closed circles indicate the oneswith available spectroscopic redshifts. The QSOs are atRA(J2000) = 61 . - . . . R band and atdifferent angular distances ∆ θ to the QSO. The results arepresented in the top panels of Figure 6. We find that oursurvey is most complete in the field around HE 0226 - R =23 at angular distances ∆ θ ≤ ′ . For PKS 0405 -
123 andPG 1216 + R = 20 (solid lines), >
80% at R = 21(dotted lines) and >
60% at R = 22 (dash-dotted lines) in theinner 2 ′ radius. The survey completeness becomes ≈
50% atlarger radii.The bottom panels of Figure 6 show the redshift distribu-tions of the spectroscopically identified galaxies in the threefields. The solid curve in each panel represents the model ex-pectation from convolving a non-evolving rest-frame R -bandgalaxy luminosity function of Blanton et al. (2003), which ischaracterized by M R ∗ - h = - . α = - .
1, and φ ∗ =0 . h Mpc - , with the completeness functions displayed inthe top panels. Redshift spikes indicate the presence of large-scale galaxy overdensities along the QSO lines of sight. Thecomparison between observations and model expectations inHE 0226 - +
069 further demonstrates thatthe color selection criterion described in § 3.2 and Figure 3has effectively excluded most galaxies at z > .
5. CATALOGS OF QSO ABSORPTION-LINE SYSTEMS
A necessary component of the galaxy–absorber cross-correlation analysis is an absorber catalog. A catalog of 57Ly α absorbers along the sightline toward HE 0226 - α absorbersexhibit a range of H I column density from log N (H I) = 12 . N (H I) = 15 . z Ly α = 0 . - .
4. We adopt their cat-alog in our cross-correlation analysis in this field. To summa-rize, the left panels of Figure 7 display the redshift distribution(top) and the cumulative N (H I) distribution function (bottom)of these absorbers.The sightline toward PKS 0405 -
123 has been studied byWilliger et al. (2006) and Lehner et al. (2007). We adopt therevised catalog of 76 Ly α absorbers from Lehner et al. (2007)in our analysis. These Ly α absorbers exhibit a range of H Icolumn density from log N (H I) = 12 . N (H I) = 16 . z Ly α = 0 . - . N (H I) distri-bution.A systematic survey of Ly α absorbers for the sightline to-ward PG 1216 +
069 has only been conducted for strong sys-tems by Jannuzi et al. (1998) in low-resolution (FWHM ≈
250 km s - ) spectra obtained using the Faint Object Spectro-graph. These authors identified nine strong Ly α absorberswith log N (H I) >
14. To complete the study of low-columndensity gas (log N (H I) <
14) along the sightline, we have con-ducted our own search of intervening Ly α lines in availableechelle spectra of the QSO obtained by STIS.Details regarding the reduction and processing of the STISechelle spectra can be found in Thom & Chen (2008a). Thefinal combined spectrum of PG 1216 +
069 covers a spectralrange of λ = 1160 - ≈ . - and has SNR of ≈ α absorbers, we first identify absorption features that areat > σ level of significance. Next, we identify known fea-tures due to either interstellar absorption of the Milky Way, ormetal absorption lines and higher order Lyman series associ-ated with known strong Ly α absorbers. We consider the re-maining unidentified absorption lines as intervening Ly α ab-sorbers. Finally, we perform a Voigt profile analysis usingthe VPFIT package for estimating the underlying N (H I) andthe Doppler parameter b (e.g. Carswell et al. 1991). This pro-cedure has yielded 66 Ly α absorbers of H I column densitylog N (H I) = 12 . - . z Ly α = 0 . - . N (H I) and the associated error, and b . The redshift distribution and cumulative N (H I) distribu-tion of these Ly α absorbers are presented in the right panelsof Figure 7.For O VI absorbers along the sightlines toward the threeQSOs in our survey, we adopt the catalog of Thom & Chen(2008a,b) and include absorbers at z < .
14 from Tripp et al.(2008). There are five O VI absorbers known along the sight-line toward HE 0226 - -
123 (see also Prochaska et al.2004), and two O VI absorbers known along the sightline to-ward PG 1216 + z = 0 . -
123 shows two distinct com-ponents separated by ≈
87 km s - (see Thom & Chen 2008b;Prochaska et al. 2004). The O VI absorber at z = 0 . +
069 also shows two dominantcomponents well separated by ≈
350 km s - (see § 6.3 andFigure 12 below). We consider these components separateobjects in the cross-correlation study presented in § 7. Theredshifts of these O VI absorbers range from z O VI = 0 .
017 to z O VI = 0 . N (O VI) = 13 . N (O VI) = 14 .
7. The redshiftdistribution of these O VI absorbers is shown in shaded his-tograms in the top panels of Figure 7.
6. DESCRIPTION OF INDIVIDUAL FIELDS
In this section, we review the galaxy and absorber proper-ties in each individual field.6.1.
The Field toward HE0226 - QSO = 0 . - α ab-sorbers (Lehner et al. 2006) and five O VI absorbers (Trippet al. 2008; Thom & Chen 2008a,b) over the redshift rangefrom z Ly α = 0 .
017 to z Ly α = 0 .
4. The column density ofthese Ly α absorbers span a range from log N (H I) = 12 . N (H I) = 15 .
1, and the column density of the O VI ab-sorbers span a range from log N (O VI) = 13 . N (O VI) = orrelation of Galaxies and QSO Absorption-line Systems 9 F IG . 5.— Summary of the spectroscopic survey in the fields around HE 0226 - -
123 (middle), and PG 1216 +
069 (right). Open circlesrepresent all objects of R ≤
23 in HE 0226 - R ≤
22 in PKS 0405 -
123 and PG 1216 + . - . . - . . . .
4. This field is particularly interesting because it exhibitsa Ne VIII λλ z = 0 . + (207.28 e.V.) and the relatively largecosmic abundance of Ne make the Ne VIII doublet transitionsa sensitive probe of warm-hot gas of T = (0 . - × K.Savage et al. (2005) performed a detailed analysis of the ion-ization state of the gas, taking into account the relative abu-dances of additional highly ionized species such as S VI andO VI, and concluded that the observations are best explainedby a collisional ionized gas of T = 5 . × K. While thedetection of Ne VIII at z = 0 .
207 provides exciting support forthe presence of warm-hot gas, this system remains unique andcontinuing efforts to search for more Ne VIII absorbers haveuncovered no new systems (e.g. Lehner et al. 2006).Our galaxy survey is most complete in the field around HE 0226 - R = 23 at angular distances ∆ θ ≤ ′ (upper-left panel of Figure 6). The redshift distribution shown inthe bottom-left panel of Figure 6 displays clear galaxy over-densities at z = 0 .
27 and z = 0 . α absorbers present at these redshifts, no O VIabsorbers are found to coincide with these large-scale galaxyoverdensities to the limit of log N (O VI) = 13 . . ′ × . ′ field around the QSO. We have been able to obtain high S / N spectra of galaxies at angular distance as close as ∆ θ . ′′ to the background bright QSO. Galaxies with redshift co-incide with known O VI absorbers ( | ∆ v | <
300 km s - ) aremarked by a rectangular box. A complete photometric andspectroscopic catalog of galaxies with R ≤
23 and at angulardistance ∆ θ ≤ ′ of the QSO is available electronically athttp://lambda.uchicago.edu/public/cat/cat_0226.html. An ex-0 Chen & Mulchaey F IG . 6.— Summary of the survey completeness (top) and redshift distributions (bottom) of galaxies in the fields around HE 0226 - - +
069 (right). The survey completeness was calculated for different brightness limits, from R ≤
20 (solid lines), to R ≤
21 (dotted lines)and to R ≤
22 (dash-dotted lines). The bottom panels show the redshift distributions of spectroscopically identified galaxies in the three fields, in comparison tomodel expectations (solid curves) based on a non-evolving rest-frame R -band galaxy luminosity function of Blanton et al. (2003) and the respective completenessfunctions displayed in the top panels. The redshift distribution of absorption-line dominated galaxies in each field is shown in solid histograms. Large-scalegalaxy overdensities along the QSO lines of sight are evident through the presence of redshift spikes. ample of the first 30 targets in the catalog is presented in Table4, which lists from columns (1) through (13) the object ID, theright ascension (RA) and declination (Dec), the position off-sets in RA ( ∆ α ) and Dec ( ∆ δ ) of the galaxy from the QSO,the angular distance to the QSO ( ∆ θ ), the projected distancein physical h - kpc, the BRI magnitudes and uncertainties, thespectroscopic redshift z spec ( - R -band absolute magnitude .Of the five O VI absorbers found along the sightline to-ward HE 0226 - | ∆ v | <
300 km s - and projected physicaldistances ρ ≤ h - kpc for two absorbers, including the The rest-frame R -band magnitude of each spectroscopically identifiedobject was estimated based on the observed R -band magnitude and spectraltype. For absorption-line dominated galaxies, we evaluate the k -corretion us-ing the early-type E/S0 and Sab galaxy templates from Coleman et al. (1980).For emission-line dominated galaxies, we evaluate the k -correction using theIrr templates. Ne VIII absorber at z = 0 . z = 0 . ρ < h - physical kpc at this low redshift.Three galaxies of R = 20 . - . | ∆ v | <
300 km s - and ρ < h - physical kpc of the O VI andNe VIII absorber at z = 0 . L ∗ galax-ies with rest-frame R -band absolute magnitudes spanning arange from M R - h = - . M R - h = - .
9. Thesurvey completness rules out the presence of additional galax-ies that are more luminous than M R - h = - . ρ = 285 h - kpc. A detailed analysis of the galactic envi-ronment of the Ne VIII absorber is presented in Mulchaey &Chen (2009).The nearest galaxy to the O VI absorber at z = 0 . R = 23 . ρ = 381 h - phys-ical kpc, with a corresponding rest-frame R -band absolutemagnitude of M R - h = - .
7. The galaxy spectrum isdominated by absorption features. Three more galaxies areorrelation of Galaxies and QSO Absorption-line Systems 11 F IG . 7.— The redshift distribution (top) and cumulative N (HI) distribution (bottom) of Ly α absorbers identified along the sightline toward HE 0226 - -
123 (middle), and PG 1216 +
069 (right). The shaded histograms in the top panels show the redshift distributions of OVI absorbers along therespective sightlines. found at | ∆ v | <
300 km s - from the absorber redshift, butthey are over ρ = 840 h - kpc to ρ = 2 . h - Mpc physical dis-tances away. The survey completness rules out the presence ofadditional star-forming galaxies that are more luminous than M R - h = - . ρ = 396 h - kpc.The nearest galaxy to the O VI absorber at z = 0 . R = 22 . ρ = 213 h - physical kpc, with a correspond-ing rest-frame R -band absolute magnitude of M R - h = - .
3. The galaxy spectrum is dominated by emission fea-tures. Two more galaxies are found at | ∆ v | <
300 km s - from the absorber redshift, and have ρ = 432 h - kpc and ρ = 643 h - kpc physical distances, respectively. The sur-vey completness rules out the presence of additional galax-ies that are more luminous than M R - h = - . ρ = 407 h - kpc.The nearest galaxy to the O VI absorber at z = 0 . R = 22 . ρ = 306 h - physical kpc, with a correspond-ing rest-frame R -band absolute magnitude of M R - h = - .
5. The galaxy spectrum is dominated by emission fea- tures. Ten more galaxies are found at | ∆ v | <
300 km s - from the absorber redshift over a physical distance range of ρ = 387 h - kpc to ρ = 1 . h - Mpc. The survey completnessrules out the presence of additional galaxies that are more lu-minous than M R - h = - . ρ = 419 h - kpc.6.2. The Field toward PKS0405 -
123 at z
QSO = 0 . -
123 exhibits 76 Ly α absorbers (Lehner et al. 2006) and six O VI absorbers(Prochaska et al. 2004; Tripp et al. 2008; Thom & Chen2008a,b) over the redshift range from z Ly α = 0 .
012 to z Ly α =0 . α absorbers span arange from log N (H I) = 12 . N (H I) = 16 .
3, and thecolumn density of the O VI absorbers span a range fromlog N (O VI) = 13 . N (O VI) = 14 . -
123 has been surveyed byProchaska et al. (2006). This previous survey has yielded ro-bust redshifts for 95% of all galaxies brighter than R = 20,corresponding to roughly L ∗ galaxies at z = 0 .
5. Our programexpands upon the earlier spectroscopic survey for uncoveringfainter galaxies along the QSO line of sight, reaching 100%2 Chen & Mulchaey
TABLE 3IGM Ly α A BSORBERS I DENTIFIED A LONG THE S IGHTLINE TOWARD
PG 1216 + z Ly α log N (HI) b ( kms - ) z Ly α log N (HI) b ( kms - )(1) (2) (3) (4) (5) (6)0.00362 13 . ± .
17 49 ±
24 | 0.13503 b . ± .
04 35 ± a , b . ± .
03 ... | 0.15475 13 . ± .
09 29 ± . ± .
03 40 ± . ± .
16 24 ± . ± .
10 34 ± . ± .
13 18 ± . ± .
14 59 ±
22 | 0.15601 13 . ± .
07 14 ± . ± .
18 20 ±
10 | 0.16076 13 . ± .
10 67 ± . ± .
03 26 ± b . ± .
03 34 ± . ± .
12 35 ±
12 | 0.18026 13 . ± .
10 34 ± . ± .
11 25 ± . ± .
04 33 ± . ± .
12 23 ± . ± .
06 40 ± . ± .
03 35 ± . ± .
11 23 ± . ± .
08 31 ± b . ± .
03 93 ± . ± .
24 45 ±
11 | 0.20114 13 . ± .
09 23 ± b . ± .
15 32 ± . ± .
12 25 ± . ± .
19 16 ± . ± .
06 42 ± . ± .
16 35 ±
16 | 0.22578 12 . ± .
13 22 ± . ± .
12 16 ± . ± .
10 32 ± . ± .
07 37 ± . ± .
18 17 ± . ± .
18 64 ±
32 | 0.23641 13 . ± .
12 43 ± . ± .
07 54 ±
10 | 0.25035 12 . ± .
14 17 ± . ± .
08 30 ± . ± .
12 26 ± . ± .
10 40 ±
12 | 0.26624 12 . ± .
14 28 ± . ± .
03 39 ± b . ± .
04 32 ± . ± .
20 44 ±
26 | 0.27154 12 . ± .
12 12 ± . ± .
05 20 ± . ± .
07 22 ± . ± .
15 78 ±
32 | 0.27865 12 . ± .
15 42 ± . ± .
15 15 ± . ± .
18 16 ± . ± .
08 32 ± b . ± .
10 21 ± . ± .
08 29 ± . ± .
19 21 ± b . ± .
14 28 ± . ± .
09 78 ± b . ± .
02 53 ± . ± .
13 47 ± . ± .
07 20 ± . ± .
11 79 ± . ± .
09 20 ± . ± .
12 21 ± a The measurement is adopted from Tripp et al. (2005). b These lines have also been identified by Jannuzi et al. (1998) in low-resolutiondata obtained using the Faint Object Spectrograph. completeness at R = 20 and >
70% at R = 22 within ∆ θ ≤ ′ (upper-middle panel of Figure 6). The redshift distributionshown in the bottom-middle panel of Figure 6 displays nu-merous galaxy overdensities in front of the QSO, but only oneO VI absorber is found to coincide with the large-scale galaxyoverdensity at z = 0 . - . ′ × . ′ area. Galaxieswith known spectroscopic redshifts are indicated by their red-shifts to the left. Blue values represent redshift measurementsobtained by previous authors (Spinrad et al. 1993; Prochaskaet al. 2006). Red values represent new redshifts obtained inour survey. Galaxies with redshift coincident with knownO VI absorbers ( | ∆ v | <
300 km s - ) are marked by a rectan-gular box. While this field contains a relatively high surfacedensity of galaxies, the majority of the galaxies turn out toreside in the QSO host environment. A complete photomet-ric and spectroscopic catalog of galaxies with R ≤
22 and atangular distance ∆ θ ≤ ′ of the QSO is available electroni-cally at http://lambda.uchicago.edu/public/cat/cat_0405.html.An example of the first 30 targets in the catalog is presentedin Table 5.Of the six O VI absorbers found along the sightline towardPKS 0405 - | ∆ v | <
300 km s - and ρ ≤ h - kpc. For the O VI ab-sorber at z = 0 . ρ = 306 h - kpc with R = 19 . ∆ v = -
341 km s - (Prochaska et al. 2006). We have iden-tified three new galaxies of R = 21 . - . | ∆ v | = 27 - - and ρ = 74 - h - kpc from the absorber, includ-ing one that is observed in the combined HST/WFPC2 im-age (left panel of Figure 10). All three galaxies are faintdwarfs with emission-line dominated spectral features (seethe top three panels of Figure 15 below). The correspond-ing rest-frame R -band absolute magnitudes span a range from M R - h - = - . M R - h - = - .
1. The HSTimage presented in Figure 10 shows that the dwarf galaxy at ρ = 73 . h - kpc exhibits a compact core with low-surfacebrightness emission extended to the north of the galaxy.Only one galaxy is found within | ∆ v | <
300 km s - and ρ < h - physical kpc of the O VI absorber at z = 0 . R = 19 .
0, correspond-ing to M R - h = - . | ∆ v | <
300 km s - from theabsorber redshift and between ρ = 257 h - kpc to ρ = 945 h - kpc physical distances away.The strong O VI absorber of log N (O VI) = 14 . z =0 . F IG . 8.— A summary of the spectroscopic survey of faint galaxies in the center 2 . ′ × . ′ field around HE 0226 - z QSO = 0 . B , R , and I images from our preimaging data. We have reached 100% completeness for galaxies brighter than R = 23 in this area.Our best estimated redshifts are shown to the left of individual sources. Galaxies with redshift coincident with known O VI absorbers ( | ∆ v | <
300 kms - ) aremarked by a rectangular box. Most interestingly, at the location of the Ne VIII absorbers, z = 0 . ρ < h - kpc . (2000) and Prochaska et al. (2004). Two galaxies of R = 17 . R = 21 . | ∆ v | <
200 km s - and ρ < h - physical kpc of the O VI absorber at z = 0 . ρ =1 h - Mpc of the absorber. The luminous galaxy at ∆ δ =40 ′′ (corresponding to ρ = 80 . h - kpc) from the QSO has M R - h = - . ρ = 67 . h - kpc has M R - h = - .
1. The disk-like morphology appears to bemildly distburbed by a faint companion at the northeast edge(center-right panel of Figure 10).No galaxies are found near the O VI absorber at z = 0 . R = 22 within ∆ θ = 2 ′ suggests that the galaxies associated with the O VI ab-sorber is likely fainter than M R - h = - . z = 0 . R = 22 within ∆ θ = 2 ′ suggests that the galaxies associated with the O VI ab-sorber is likely fainter than M R - h = - . N (O VI) = 14 . z =0 . ρ = 2 . h - Mpc. Our spectroscopicsurvey has uncovered a galaxy of R = 22 .
63 at ∆ v = - - and ρ = 77 h - physical kpc from the absorber, witha corresponding rest-frame R -band absolute magnitude of M R - h = - .
8. The spectrum is dominated by emission-line features due to [O II], H γ , H β , and [O III] (see the 4thpanel in Figure 15 below). The morphology appears to beextended in the HST image in Figure 10 (right panel). No ad-ditional galaxies are found at | ∆ v | <
300 km s - and ρ < h - TABLE 4A N E XAMPLE OF THE P HOTOMETRIC AND S PECTROSCOPIC C ATALOG OF O BJECTS IN THE F IELD AROUND
HE0226 - a ∆ α ∆ δ ∆ θ ρ M R ID RA(J2000) Dec(J2000) ( ′′ ) ( ′′ ) ( ′′ ) ( h - kpc) B R I z spec
Type b - h (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)00443 02:29:21.173 - . . . - .
00 22 . ± .
034 20 . ± .
006 18 . ± . - . . - . . . - .
00 23 . ± .
066 22 . ± .
024 21 . ± . - . . - . . . - .
00 23 . ± .
048 22 . ± .
022 21 . ± . - . . - . . . - .
00 25 . ± .
200 22 . ± .
031 21 . ± . - . . - . . . - .
00 22 . ± .
024 20 . ± .
006 19 . ± . - . . - . - . . - .
00 23 . ± .
066 21 . ± .
023 20 . ± . - . . - . - . . - .
00 24 . ± .
147 22 . ± .
048 22 . ± . - . . - . . . - .
00 22 . ± .
035 19 . ± .
004 18 . ± . - . . - . . . .
79 22 . ± .
021 20 . ± .
009 19 . ± .
008 0 . - . - . . . .
03 23 . ± .
047 21 . ± .
021 21 . ± .
021 0 . - . - . - . . - .
00 25 . ± .
267 22 . ± .
035 21 . ± . - . . - . - . . .
43 23 . ± .
051 21 . ± .
014 20 . ± .
012 0 . - . - . . . - .
00 24 . ± .
160 22 . ± .
038 22 . ± . - . . - . . . - .
00 24 . ± .
137 22 . ± .
045 22 . ± . - . . - . - . . - .
00 22 . ± .
041 19 . ± .
005 18 . ± . - . . - . . . - .
00 24 . ± .
220 22 . ± .
031 21 . ± . - . . - . . . .
70 21 . ± .
016 19 . ± .
003 18 . ± .
002 0 . - . - . . . - .
00 24 . ± .
137 21 . ± .
021 20 . ± . - . . - . - . . - .
00 24 . ± .
197 21 . ± .
020 20 . ± . - . . - . . . - .
00 23 . ± .
060 22 . ± .
029 21 . ± . - . . - . - . . .
54 23 . ± .
060 21 . ± .
020 20 . ± .
016 0 . - . - . . . - .
00 24 . ± .
095 22 . ± .
036 21 . ± . - . . - . - . . - .
00 24 . ± .
095 22 . ± .
049 21 . ± . - . . - . . . - .
00 24 . ± .
172 22 . ± .
046 21 . ± . - . . - . - . . - .
00 25 . ± .
182 22 . ± .
042 21 . ± . - . . - . . . - .
00 24 . ± .
099 22 . ± .
041 22 . ± . - . . - . . . - .
00 24 . ± .
081 22 . ± .
034 21 . ± . - . . - . - . . - .
00 22 . ± .
037 20 . ± .
015 20 . ± . - . . - . . . - .
00 23 . ± .
076 22 . ± .
035 21 . ± . - . . - . . . - .
00 25 . ± .
282 22 . ± .
038 21 . ± . - . . a A complete catalog is available electronically at http://lambda.uchicago.edu/public/cat/cat_0226.html. b Spectral type of the galaxy: 1 → absorption-linedominated and 2 → emission-line dominated. Mpc.6.3.
The Field toward PG1216 +
069 at z
QSO = 0 . α absorbers described in § 5 has identified66 Ly α absorbers along the sightline toward PG 1216 + z Ly α = 0 . z Ly α = 0 . α absorbersspan a range from log N (H I) = 12 . N (H I) = 19 . z = 0 . z = 0 . N (O VI) = 14 . N (O VI) = 13 .
4, respectively(Tripp et al. 2008; Thom & Chen 2008a,b) .Our galaxy survey in the field around PG 1216 +
069 in theleast incomplete of all three, reaching >
80% completenessfor galaxies brighter than R = 21 at angular distances ∆ θ ≤ ′ (upper-right panel of Figure 6). The redshift distributionshown in the bottom-right panel of Figure 6 displays a sig-nificant galaxy overdensity at z = 0 . - .
28 in front of theQSO. One of the two O VI absorbers does coincide with thislarge-scale galaxy overdensity. Figure 11 shows the redshiftsof galaxies in the inner 2 . ′ × . ′ field around the QSO. Wehave not been able to obtain high quality spectra of galaxiesat angular distance ∆ θ . ′′ due to the presence of a bright We note that Tripp et al. (2008) reported one more OVI absorber at z = 0 . β featureat z = 0 . α at z = 0 . star and the background QSO. Only one foreground galaxyis found at ρ < h - kpc. A complete photometric andspectroscopic catalog of galaxies with R ≤
22 and at angulardistance ∆ θ ≤ ′ of the QSO is available electronically athttp://lambda.uchicago.edu/public/cat/cat_1216.html. An ex-ample of the first 30 targets in the catalog is presented in Table6. A galaxy is identified at | ∆ v | <
300 km s - and ρ ≤ h - kpc from the absorber at z = 0 . R = 19 .
3, corresponding to M R - h = - .
0, and ρ = 64 h - kpc. The morphology of the galaxy isindicative of a galaxy (at the southern edge) being tidally tornby a more massive galaxy during the process of merging (top-left panel of Figure 12). The spectrum of the system shownin the bottom left panel of Figure 12 exhibits prominent emis-sion features, such as H α , [N II], and [S II], suggesting thatthe ISM is chemically enriched to solar metallicity. The largeH α to H β line ratio also suggests the presence of substantialdust in the ISM.The absorber at z = 0 . δ v ≈ . - ; the combined FUSE spectrum has aspectral resolution of δ v ≈
30 km s - .orrelation of Galaxies and QSO Absorption-line Systems 15 F IG . 9.— A combined HST image obtained using WFPC2 and the F702W filter. The image is roughly 2 . ′ on a side. Galaxies with known spectroscopicredshifts are indicated by their redshifts to the left. Blue values represent redshift measurements obtained by previous authors (Spinrad et al. 1993; Prochaskaet al. 2006). Red values represent new redshifts obtained in our survey. Similar to Figure 8, galaxies with redshift coincident with known O VI absorbers( | ∆ v | <
300 kms - ) are marked by a rectangular box. At least four absorption components ( ∆ v = - , - , + +
188 km s - ) are seen in the H I, C III, and O VI absorp-tion transitions (see also Tripp et al. 2008 for a brief discus-sion of this system). Narrow absorption due to the Si III λ α absorption are saturated. Wehave performed a Voigt profile analysis to determine the col-umn densities of individual components. The results are pre-sented in Table 7, which lists from columns (2) through (9)the best-fit column density, Doppler parameter ( b ), and theirassociated uncertainties. Column (10) of Table 7 gives the to-tal estimated column densities of individual transitions. Giventhe saturated profiles of the Ly α transition and relatively lowresolution in the observation of Ly β , we place conservativelimits for the underlying N (H I). The Si III components allappear to be very narrow and unresolved in the STIS echelledata. We estimate the column densities based on a fixed b value of b = 2 . - that corresponds to gas of ≈ K.The best-fit Voigt profiles are shown in the right panel of Fig-ure 12 as smooth curves. While the blue-shifted and redshifted components of Ly α and Ly β show nearly symmetric kinematic signatures (andpossibly in the C III absorption as well) that are indicativeof an origin in expanding supperbubble shells (see e.g. Bondet al. 2001 and Simcoe et al. 2002, although see also Kawata& Rauch 2007 who showed that such absorption featurescould also be produced by filamentary accretion onto a centralgalaxy or galaxy group), there exists a strong abundance gra-dient in Si III and O VI from ∆ v = -
158 km s - to ∆ v = + - that suggests a significant variation in the underlyinggas density (assuming a uniform background radiation field).Combining the optical morphology of the galaxy and the dif-ferential relative abundances of different ionization species in-dicates that the absorber is likely to originate in disrupted tidaltails as a result of a merger.The nearest galaxy to the O VI absorber at z = 0 . R = 19 . ρ = 374 h - physical kpc, with a corresponding rest-frame R -band absolute magnitude of M R - h = - .
3. The galaxyspectrum is dominated by absorption features. Three more6 Chen & Mulchaey F IG . 10.— Optical morphologies of galaxies that are associated with known OVI absorbers along the sightline toward PKS 0405 - h - physical kpc on a side at the redshift of the galaxy. The arrow in each panel indicates the direction toward the QSO sightline. galaxies are found at | ∆ v | <
300 km s - from the absorberredshift, but they are over ρ = 835 h - kpc to ρ = 2 . h - Mpcphysical distances away.
7. ANALYSIS
We have obtained a spectroscopic sample of galaxies infields around three QSOs, for which ultraviolet echelle spec-tra from FUSE and HST/STIS are available for identifyingintervening hydrogen Ly α and O VI absorbers. The galax-ies span a broad range in the rest-frame R -band magnitudefrom M R - h > -
16 to M R - h < -
22 and a broadrange in the projected physical distance from ρ < h - kpcto ρ > h - Mpc. We have shown in § 4.2 that the com-pleteness of the spectroscopic survey is well understood andcharacterized by the angular selection function presented inFigure 6. Together with a complete sample of interveningLy α and O VI absorbers found at z = 0 . - . α and O VI absorbers basedon their cross-correlation amplitude with known galaxies, andfor investigating the gas content in halos around galaxies. 7.1. The Galaxy–Absorber Cross-Correlation Functions
To quantify the origin of Ly α and O VI absorbers, we firstmeasure the projected two-point galaxy auto-correlation func-tion using a flux-limited sample that has been assembled fromour spectroscopic survey. The flux-limited galaxy samplecontains 670 spectroscopically identified intervening galax-ies of R ≤
22 within ∆ θ = 11 ′ of the background QSOs, 222of which show absorption-line dominated spectral features.The median rest-frame R -band absolute magnitude of the 448emission-line dominated galaxies is h M R i - h = - . ≈ . L ∗ for M R ∗ - h = - .
44 fromBlanton et al. 2003), while the 222 absorption-line domi-nated galaxies have h M R i - h = - .
43 (corresponding to ≈ L ∗ ). We calculate the projected two-point correlation func-tion ω gg ( r p ) versus co-moving projected distance r p , using theLandy & Szalay (1993) estimator ω gg ( r p ) = D g D g - D g R g + R g R g R g R g , (1)where D g represents the input galaxy sample and R g repre-sents a random galaxy sample. The random galaxy sampleorrelation of Galaxies and QSO Absorption-line Systems 17 TABLE 5A N E XAMPLE OF THE P HOTOMETRIC AND S PECTROSCOPIC C ATALOG OF O BJECTS IN THE F IELD AROUND
PKS0405 - a ∆ α ∆ δ ∆ θ ρ M R ID RA(J2000) Dec(J2000) ( ′′ ) ( ′′ ) ( ′′ ) ( h - kpc) R z spec
Type b - h (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)00071 04:08:33.10 - . . - .
00 21 . ± .
110 -1.0000 0 0 . c - . - . .
30 20 . ± .
090 0.5639 1 - . - . - . - .
00 20 . ± .
080 -1.0000 0 0 . - . - . - .
00 19 . ± .
070 -1.0000 0 0 . - . . - .
00 21 . ± .
100 -1.0000 0 0 . - . . - .
00 21 . ± .
090 -1.0000 0 0 . - . . - .
00 16 . ± .
060 -1.0000 0 0 . - . - . - .
00 18 . ± .
060 -1.0000 0 0 . - . - . - .
00 20 . ± .
090 -1.0000 0 0 . - . . - .
00 17 . ± .
060 -1.0000 0 0 . c - . - . .
03 20 . ± .
090 0.5645 2 - . - . - . - .
00 21 . ± .
090 -1.0000 0 0 . - . . - .
00 17 . ± .
060 -1.0000 0 0 . - . - . - .
00 21 . ± .
120 -1.0000 0 0 . - . - . - .
00 21 . ± .
110 -1.0000 0 0 . - . . - .
00 21 . ± .
130 -1.0000 0 0 . - . - . - .
00 21 . ± .
100 -1.0000 0 0 . - . . - .
00 17 . ± .
060 -1.0000 0 0 . - . - . - .
00 16 . ± .
060 -1.0000 0 0 . - . . - .
00 19 . ± .
070 -1.0000 0 0 . - . . - .
00 21 . ± .
090 -1.0000 0 0 . - . . - .
00 21 . ± .
110 -1.0000 0 0 . - . . - .
00 21 . ± .
120 -1.0000 0 0 . - . - . - .
00 21 . ± .
130 -1.0000 0 0 . - . - . - .
00 20 . ± .
080 -1.0000 0 0 . - . . - .
00 19 . ± .
070 -1.0000 0 0 . - . . - .
00 20 . ± .
090 -1.0000 0 0 . - . - . - .
00 21 . ± .
100 -1.0000 0 0 . - . . - .
00 21 . ± .
110 -1.0000 0 0 . d - . - . .
39 19 . ± .
070 0.3466 1 - . a A complete catalog is available electronically at http://lambda.uchicago.edu/public/cat/cat_0405.html.The Object ID, coordinates, and R -band photometry are adopted from Prochaska et al. (2006). b Spectral type of the galaxy: 1 → absorption-linedominated and 2 → emission-line dominated. c New Spectroscopic redshifts obtained in our survey. d Spectroscopic redshifts published in Prochaska et al. (2006). is generated following the completeness function presentedin Figure 6 for each field. To minimize possible countingnoise introduced by the random galaxy sample, we have pro-duced a random galaxy sample per field that is a factor of tenlarger than the true galaxy sample. The two-point function isthen calculated by counting the appropriate D g D g , D g R g , and R g R g pairs within different r p intervals, from r p < h - kpc, to r p = 250 - h - kpc, to r p = 800 - h - kpc, andto r p = 1 . - h - Mpc.We present in panel (a) of Figure 13 the auto-correlationfunctions intergrated over 40 h - co-moving Mpc in redshiftspace for all galaxies (open circles), absorption-line domi-nated galaxies (open triangles), and emission-line dominatedgalaxies (open squares) in the flux-limited sample. Error-bars represent poisson counting uncertainties . These datapoints for each of these subsamples are also repeated in pan-els (b), (c), and (d), respectively, to be compared with theircorresponding galaxy–absorber cross-correlation functions.Panel (a) of Figure 13 confirms that absorption-line domi- We note that on small scales ( r p . h - Mpc) the correlation ampli-tude is dominated by satellite galaxies (e.g. Zheng et al. 2007). Errors in themeasured clustering signals on these scales are expected to be roughly pois-son counting errors. On larger scales, however, field-to-field variations areexpected to dominate the errors in the observed clustering signals and the re-ported poisson errors represent a lower limit to the true uncertainties. Giventhat our survey covers only three QSO fields, we are unable to evaluate theuncertainties due to field-to-field variations using a jackknife method. nated galaxies indeed cluster more strongly than emission-line dominated galaxies. The observed clustering amplitudeof absorption-line dominated galaxies in our sample is com-parable to what is found by Zehavi et al. (2005) for SDSSgalaxies of M R - h < -
19 at z ∼ .
1. The stronger cluster-ing amplitude indicates that on average these absorption-linegalaxies originate in higher overdensity regions and presum-ably more massive dark matter halos.Next, we measure the projected two-point galaxy–absorbercross-correlation function using the flux-limited galaxy sam-ple discussed above and the absorber catalogs discussed in §5. We exclude absorbers that are within ∆ v = 3000 km s - ofthe background QSO to reduce contaminations due to QSOassociated absorbers. There are 195 Ly α absorbers and 15O VI absorbers included in our analysis. We calculate the pro-jected two-point cross-correlation function ω ga ( r p ) versus co-moving projected distance r p between galaxies and absorbers,using the Landy & Szalay (1993) estimator ω ga ( r p ) = D g D a - D g R a - R g D a + R g R a R g R a , (2)where D g and D a represent the input galaxy and absorber sam-ples, and R g and R a represent random galaxy and absorbersamples. We have produced random Ly α and O VI absorbercatalogs for each field, assuming a flat absorber selectionfunction in the redshift interval between z = 0 .
01 and z QSO .The random absorber catalogs are also ten times larger than8 Chen & Mulchaey F IG . 11.— A combined HST image obtained using WFPC2 and the F702W filter. The image is roughly 2 . ′ on a side. Galaxies with known spectroscopicredshifts are indicated by their redshifts to the left. Known redshifts obtained prior to our survey are marked in blue (Chen et al. 2001a). Red values representour own measurements. The galaxy with a redshift coincident with the O VI absorber at z = 0 . ρ < h - kpc. the true absorber catalogs. The two-point cross-correlationfunction is then calculated by counting the appropriate D g D a , D g R a , R g D a , and R g R a pairs within the same r p intervals ofthe auto-correlation function calculation.Panels (b), (c), and (d) of Figure 13 show the galaxy–Ly α absorber and galaxy–O VI absorber cross-correlation func-tions intergrated over 40 h - co-moving Mpc in redshift space.Comparisons between galaxy–absorber cross-correlation andgalaxy auto-correlation functions of different galaxy typeshow four interesting features.First, while both strong and weak Ly α absorbers exhibiton average weaker clustering amplitudes than the galaxies asa whole (solid points in panel b), strong Ly α absorbers oflog N (H I) ≥
14 appear to exhibit a comparable clustering am-plitude on large scales of r p = 0 . - h - Mpc and a higherclustering amplitude on small scales of r p ≤ h - kpc withthe emission-line dominated galaxies (solid points with soliderrorbars panel c). The comparable clustering signal on largescales indicates that strong Ly α absorbers and emission-linegalaxies share common halos. The higher galaxy–absorber cross-correlation signal relative galaxy auto-correlation sig-nal on small scales is understood if the gas covering fraction ishigh in halos around these galaxies. Taking into account pre-vious findings of Chen et al. (1998; 2001a), we conclude thatemission-line dominated galaxies are surrounded by extendedgas of nearly unity covering fraction and that strong Ly α ab-sorbers primarily probe extended gaseous halos around youngstar-forming galaxies.Second, these strong Ly α absorbers exhibit on average ≈ r p < h - Mpc. The weaker clustering amplitudebetween strong Ly α absorbers and absorption-line dominatedgalaxies is qualitatively consistent with the clustering ampli-tude of Mg II absorbers and luminous red galaxies (LRGs)found by Gauthier et al. (2009) using a flux-limited LRG sam-ple. It suggests that the incidence of strong Ly α absorbers isvery low around these absorption-line dominated galaxies.Third, weak Ly α absorbers of log N (H I) < . r p < h - Mpc (solid points with dotted errorbarsorrelation of Galaxies and QSO Absorption-line Systems 19
TABLE 6A N E XAMPLE OF THE P HOTOMETRIC AND S PECTROSCOPIC C ATALOG OF O BJECTS IN THE F IELD AROUND
PG1216 + a ∆ α ∆ δ ∆ θ ρ M R ID RA(J2000) Dec(J2000) ( ′′ ) ( ′′ ) ( ′′ ) ( h - kpc) B R I z spec
Type b - h (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)00539 12:18:48.159 + - . - . . - .
00 18 . ± .
023 16 . ± .
005 16 . ± . - . . + . - . . .
15 21 . ± .
014 20 . ± .
005 19 . ± .
006 0.2943 2 - . + . - . . - .
00 21 . ± .
012 18 . ± .
002 17 . ± . - . . + . - . . - .
00 22 . ± .
019 21 . ± .
015 21 . ± . - . . + - . - . . .
46 22 . ± .
026 21 . ± .
010 20 . ± .
012 0.4328 2 - . + - . - . . - .
00 23 . ± .
046 20 . ± .
006 19 . ± . - . . + . - . . .
74 20 . ± .
015 19 . ± .
004 18 . ± .
004 0.1753 2 - . + - . - . . - .
00 22 . ± .
040 21 . ± .
019 21 . ± . - . . + . - . . .
82 21 . ± .
011 19 . ± .
004 19 . ± .
004 0.3054 2 - . + . - . . - .
00 22 . ± .
038 20 . ± .
009 19 . ± . - . . + . - . . - .
00 18 . ± .
002 17 . ± .
001 17 . ± . - . . + . - . . - .
00 22 . ± .
034 20 . ± .
006 19 . ± . - . . + - . - . . - .
00 22 . ± .
029 21 . ± .
013 21 . ± . - . . + . - . . .
71 21 . ± .
028 20 . ± .
009 19 . ± .
010 0.2981 2 - . + - . - . . .
51 22 . ± .
034 20 . ± .
010 20 . ± .
012 0.2981 2 - . + - . - . . .
59 23 . ± .
111 20 . ± .
008 19 . ± .
006 0.5220 1 - . + . - . . - .
00 22 . ± .
035 21 . ± .
017 21 . ± . - . . + . - . . - .
00 22 . ± .
046 21 . ± .
020 21 . ± . - . . + - . - . . - .
00 23 . ± .
065 21 . ± .
015 20 . ± . - . . + . - . . - .
00 23 . ± .
101 21 . ± .
019 21 . ± . - . . + - . - . . - .
00 22 . ± .
022 21 . ± .
015 21 . ± . - . . + - . - . . .
19 22 . ± .
045 21 . ± .
016 20 . ± .
020 0.2618 2 - . + . - . . - .
00 22 . ± .
018 21 . ± .
010 20 . ± . - . . + - . - . . - .
00 22 . ± .
042 19 . ± .
004 18 . ± . - . . + - . - . . .
45 23 . ± .
064 21 . ± .
014 20 . ± .
012 0.5237 2 - . + - . - . . - .
00 21 . ± .
016 20 . ± .
008 19 . ± . - . . + . - . . - .
00 23 . ± .
052 21 . ± .
022 21 . ± . - . . + - . - . . - .
00 23 . ± .
061 21 . ± .
018 20 . ± . - . . + . - . . - .
00 23 . ± .
058 21 . ± .
020 21 . ± . - . . + - . - . . - .
00 21 . ± .
011 20 . ± .
006 20 . ± . - . . a A complete catalog is available electronically at http://lambda.uchicago.edu/public/cat/cat_1216.html. b Spectral type of the galaxies: “1” indicates absorption-linedominated galaxies, “2” indicates emission-line dominated galaxies, and “0” indicates absence of spectra.
TABLE 7C
OLUMN D ENSITY M EASUREMENTS OF T HE Ly α + OVI A
BSORBER AT z = 0 . LONG THE
PG 1216 +
069 S
IGHTLINE ∆ v = -
158 kms - ∆ v = -
82 kms - ∆ v = +
120 kms - ∆ v = +
188 kms - Transition log
N b log
N b log
N b log
N b log N tot (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)H I 1215 a > . > . > . . ± . ± > . b . ± . . . ± . . . ± . . < . . ± .
2C III 977 13 . ± . ± . ± . ± . ± . ±
10 13 . ± . ±
12 14 . ± .
2O VI1031,1037 a . ± . ± . ± . ± . ± . ±
17 14 . ± . ±
13 14 . ± . a See also Tripp et al. (2008). b The lines appear to be saturated. The reported column densities are estimated for the adopted Doppler parameter, b = 2 . -
1, that corresponds to gas of ≈
104 K. in Figure 13). The weak clustering amplitude indicates thatthe majority of these weak absorbers do not share the samedark matter halos as the galaxies. This result extends the find-ing of Grogin & Geller (1998) to . . L ∗ galaxies that weakLy α absorbers of log N (H I) < . r p ≈ h - Mpc are necessary to constrain the mean over-densities where these weak Ly α absorbers reside.Finally, O VI absorbers exhibit similar clustering ampli-tudes as strong Ly α absorbers. Specifically, they show a com-parable clustering amplitude (crosses in panel c of Figure 13)with emission-line galaxies on scales of r p = 0 . - . h - Mpc but a factor of ≈ F IG . 12.— Top Left : Optical morphology of the OVI absorbing galaxy at z = 0 . h - kpc on a side at the galaxy redshift. The arrow indicates the direction toward the QSO sightline. Bottom left : The IMACS spectrum of the galaxyin the top panel, showing prominent emission features of H α , [N II], and [S II] (marked by dotted lines) that indicate an ISM metallicity comparable to solar. Theweak H β line suggests the presence of substantial dust. The gap at 6000 Å is due to a chip gap between IMACS CCDs. Right : Absorption profiles of neutral andionic species found at z = 0 . δ v ≈ . - (Thom & Chen 2008a) and FUSE with a spectral resolution of δ v ≈
30 kms - (kindly provided by J. Scott; see also Scott et al. 2004). The1- σ error array is represented by the thin histograms above the zero level. The zero-flux level and normalized continuum level are marked by the dash-dottedlines. The At least four distint components are present in the H I, C III, and O VI absorption transitions. Narrow absorption due to the Si III λ such study to yield statistically significant results.Nevertherless, the comparable clustering amplitude amongemission-line galaxies, strong Ly α absorbers, and O VI ab-sorbers is, however, difficult to interpret, because the num-ber density of O VI absorbers is found to be n O VI ≈ - α absorbers is found to be n Ly α ≈ -
30 (Weymannet al. 1998; Dobrzycki et al. 2002). It shows that on aver-age roughly half of the strong Ly α absorbers have associatedO VI. We defer a more detailed discussion of this discrepancyto § 8.2.7.2. Incidence and Covering Fraction of O VI AbsorbingGas Around Galaxies
The galaxy sample also allows us to examine the incidenceand covering fraction of O VI absorbing gas around galax-ies. We have identified a total of 20 foreground galaxies of R ≤
22 in our spectroscopic sample that occur at co-movingprojected distances r p ≤ h - kpc from the lines of sight to-ward HE 0226 - - + r p ≤ h - kpc from the QSO sightlineand within velocity offsets of ∆ v ≤
300 km s - . The adoptedsearch volume is typical of the halo size around L ∗ galaxies.Galaxies and absorbers that occur in this small volume arelikely to share a common halo.For each galaxy in this sample, we search for correspondingO VI absorption features in the available echelle spectra ofthe QSOs and measure the rest-frame absorption equivalentorrelation of Galaxies and QSO Absorption-line Systems 21 F IG . 13.— ( a ) Projected auto-correlation functions of galaxies in our flux-limited sample. Open circles represent the auto-correlation function includingall galaxies; open triangles represent the auto-correlation function of absorption-line dominated galaxies only; and open squares represent the auto-correlationfunction of emission-line dominated galaxies only. We have applied a small offset in projected co-moving distance r p to measurements for different subsamplesfor clarity. Each of the auto-correlation functions is repeated in panels ( b ), ( c ), and ( d ), respectively, for the corresponding type for direct comparison with thegalaxy–absorber cross-correlation functions. ( b ) Projected galaxy–absorber cross-correlation functions, including all galaxies in the flux-limited sample. Thesolid points represent galaxy and Ly α absorber cross-correlation functions; points with solid errorbars are for strong Ly α absorbers of log N (HI) ≥
14; pointswith dotted errorbars are for weak Ly α absorbers of log N (HI) = 12 . - .
5. Stellar symbols with dashed errorbars represent the galaxy and OVI absorber cross-correlation function. The shaded line highlights the galaxy auto-correlation function to guide the comparisons with galaxy–absorber cross-correlation functions.( c ) Similar to panel (b), but for emission-line dominated galaxies only. (d) Similar to panels (b) and (c), but for absorption-line dominated galaxies only. width W (1031). In the absence of an absorption feature, wedetermine a 2- σ upper limit to W (1031) for the galaxy. Figure14 shows W (1031) versus r p for these galaxies of R ≤
22 inour sample. Emission-line dominated galaxies are marked bysquares and absorption-line dominated galaxies are markedby triangles. Arrows indicate the 2- σ upper limit in W (1031)for galaxies that do not have an associated O VI absorber. Forthose galaxies with neighbors, we include only the galaxy atsmallest r p . These galaxies are marked by an additional opencircle (indicating at least one neighboring galaxy is present at r p ≤ h - kpc from the absorber) in Figure 14.It is clear that none of the absorption-line dominated galax-ies in Figure 14 has a corresponding O VI absorber to a sen-sitve upper limit, consistent with the lower clustering am-plitude seen in panel (d) of Figure 13. Considering onlyemission-line dominated galaxies in our sample, we find thatnine of the 14 galaxies at r p ≤ h - kpc have an associ- ated O VI absorber. This translates to a covering fraction of κ ≈
64% for O VI absorbing gas within 250 h - co-movingprojected distance of star-forming galaxies. The coveringfraction appears to be κ ≈ r p ≤ h - kpc, asall four emission-line dominated galaxies have an associatedO VI absorber. The observed κ ≈
64% O VI covering fractionat r p ≤ h - kpc from emission-line galaxies is consistentwith the known number density statistics of strong Ly α andO VI absorbers, but it further underscores the difficulty in in-terpreting the comparable clustering measurements presentedin panel (c) of Figure 13.We note that the O VI absorbing galaxy at r p = 72 h - in Figure 14 is the complex absorber at z = 0 . + F IG . 14.— Observed OVI absorption strength for all galaxies of R ≤
22 found at z = 0 . - . r p ≤ h - kpc in the fields around HE 0226 - - + σ upper limit tothe rest-frame 1031 Å absorption equivalent width for galaxies that do nothave an associated OVI absorber. Note that there are two non-absorbinggalaxies occur at r p ≈ h - kpc. Several galaxies are found to have atleast one neighboring galaxy located within r p = 250 h - kpc and | ∆ v | ≤ - from the absorber, in which cases we include only the closest galaxy inthe plot. Points with open circles represent galaxies that either have additionalneighboring galaxies or exhibit disturbed morphologies indicative of a mergerevent. at r p < h - that does not have a corresponding O VI ab-sorber to a sensitive upper limit is an absorption-line dom-inated galaxy at z = 0 . R = 21 . r p ≈ h - and anabsorption-line dominated galaxy of R = 18 . r p = 251 h - kpc are also found from the QSO sightline. Although oursample is still small and the results have large uncertaintiesbecause of systematic bias from possible field to field varia-tions, it is interesting to find that star-forming galaxy “groups”at r p . h - kpc appear to show a higher incidence of O VIabsorbers.
8. DISCUSSION
Using a flux-limited ( R ≤
22) sample of 670 interven-ing galaxies spectroscopically identified at z < . - - + α absorbers of log N (H I) >
14 share comparable cluster-ing amplitude with emission-line dominated galaxies and thatweak Ly α absorbers of log N (H I) ≤ . ≈ α absorbers at z < . α absorbers of log N (H I) ≤ . ≈ . L ∗ galaxies. The differential clustering amplitudeof strong Ly α absorbers with different types of galaxies atprojected co-moving distances r p ≤ h - kpc further indi-cates that the incidence of strong Ly α absorbers is very lowaround absorption-line dominated galaxies.We have also studied the correlation between galaxies andO VI absorbers. While our spectroscopic survey has uncov-ered multiple large-scale galaxy overdensities in the threefields, Figures 6&7 show that only two O VI absorbers (at z = 0 . -
123 and at z = 0 . + α ab-sorbers as well). The interpretation of the clustering analysisis, however, complicated by additional observations that thecovering fraction of O VI absorbing gas is κ ≈
64% aroundemission-line dominated galaxies. While our sample is stillsmall, this low gas covering fraction is qualitatively consis-tent with the observed difference in the number densities ofstrong Ly α and O VI absorbers. It indicates that only a subsetof star-forming galaxies are associated with the observed O VIabsorption features. The discrepancy between the observedclustering amplitude and covering fraction of O VI absorbersimplies that additional variables need to be accounted for.In this section, we review the known properties of individ-ual O VI absorbing galaxies, and discuss the implications ofour results on the origin of O VI absorbers and on the extendedgaseous envelopes around galaxies.8.1. The Properties of
O VI
Absorbing Galaxies
There are 13 O VI absorption systems (15 well-separatedcomponents) identified at z = 0 . - .
495 along thesightlines toward HE 0226 - - +
069 (Tripp et al. 2008; Thom & Chen 2008a,b), in-cluding one at z = 0 .
207 with associated Ne VIII features (Sav-age et al. 2005). Our spectroscopic survey, together with pre-vious searches in these fields, has uncovered 11 galaxies atprojected co-moving distances of r p < h - kpc and ve-locity offsets of ∆ v ≤
300 km s - from six of the 13 O VIabsorbers . Although half of the O VI absorbers do not yethave an associated galaxy identified, this small sample of 11galaxies allows us to examine the common properties of O VIabsorbing galaxies, including the luminosity, galaxy environ-ment, morphology, and spectral features.First, the rest-frame R -band absolute magnitude of thegalaxies associated with six O VI absorbers span a broadrange from M R - h = - . M R - h = - . M R for the underlying absorbing galaxies of six re-maining O VI absorbers . A summary of galaxies found in In the absence of peculiar velocity field, ∆ v = ±
300 kms - would cor-respond to a co-moving pathlength of ± . h - Mpc. Adopting the luminos-ity function of Blanton et al. (2003) and the two-point galaxy auto-correlationfunction of Zehavi et al. (2005), we estimate that the probability of findinga random galaxy of L > . L ∗ at r p < h - co-moving kpc due to large-scale overdensity is ≈ r p < h - kpc. We have excluded the OVI absorber at z = 0 . orrelation of Galaxies and QSO Absorption-line Systems 23the vicinity of O VI absorbers is presented in Table 8, whichlists from colums (2) through (11) the absorber redshift z abs ,the velocity centroid relative to z abs of the O VI feature ∆ v OVI ,the O VI absorbing gas column density N (OVI), the rest-frame absorption equivalent width W rest (1031), the velocitycentroid relative to z abs of the H I feature ∆ v HI , the H I ab-sorbing gas column density N (HI), references of the absorbermeasurements, the projected co-moving distance r p and ab-solute magnitude M R of the galaxies, and references of thegalaxy measurements. We find that while the known O VI ab-sorbing galaxies exhibit a broad range of intrinsic luminosity,from < . L ∗ to ≈ L ∗ , a clear correlation between N (O VI)and M R is absent.At the same time, four of the six O VI absorbers that haveknown associated galaxies are surrounded by either a merg-ing galaxy pair or by more than one galaxy in the small vol-ume of r p < h - kpc and ∆ v ≤
300 km s - . In contrast,deep spectroscopic surveys to search for the galaxies pro-ducing absorption features of low-ionization species such asMg II λλ , .
16% of Mg II absorbersoriginating in multiple galaxy environment (e.g. Steidel et al.1997). On the other hand, not all galaxy “groups” within asimilar volume produce an associated O VI absorber in thespectrum of the background QSO. We have shown in § 7.2that a group of three galaxies at z ≈ .
268 and a group of twogalaxies at z ≈ .
199 within r p ≤ h - kpc from the sight-line toward HE 0226 - z = 0 . z = 0 . r p = 238 h - kpc from the sightline toward HE 0226 - z = 0 . z = 0 . - + Ne VIIIabsorber at z = 0 .
207 toward HE 0226 - z = 0 . + HE 0226 - ≈ h - kpc co-moving radius of the absorber. F IG . 15.— Spectra and the associated error arrays of four new OVI ab-sorbing galaxies identified in our survey (top four panels), three of whichare for the OVI absorber at z = 0 . -
123 and one for theOVI absorber at z = 0 . M R and the best-fit redshift of each galaxy are given in the upper-right corner of each panel. Dotted lines indicate the emission features foundat the best-fit galaxy redshift. For comparison, we have included in the bot-tom panel the spectrum of an absorption-line dominated galaxy at z = 0 . r p = 62 h - co-moving kpc from the sightline toward HE 0226 - M R - h = - . σ upper limit of W (1031) < .
02 Å. due to H α , [N II], and [S II] that imply roughly solar metallic-ity in the ISM. In the top four panels of Figure 15, we presentthe spectra of four additional O VI absorbing galaxies identi-fied in our survey, three of which are for the O VI absorberat z = 0 . -
123 and one for the O VI ab-sorber at z = 0 . β and [O III], are present in all four galax-ies. While the O VI absorbing galaxy at z = 0 . α emission in the three O VI absorbing galaxiesat z = 0 .
091 places 2- σ upper limits to the ISM metallicityat ≈
10% to 30% of solar values. In contrast to the spectralfeatures of O VI absorbing galaxies, we include in the bot-tom panel of Figure 15 the spectrum of an absorption-linedominated galaxy at z = 0 . r p = 62 h - co-movingkpc from the sightline toward HE 0226 - M R - h = - . σ upper limit of W (1031) < .
02 Å.Excluding the O VI absorber at z = 0 . - h - kpc, we note that four of the five strong O VI4 Chen & Mulchaey TABLE 8S
UMMARY OF G ALAXIES IN THE V ICINITY OF
OVI A
BSORBERS AT z < . a Absorbers Galaxies ∆ v OVI W rest (1031) ∆ v HI r p M R b Sightline z abs ( kms - ) log N (OVI) (mÅ) ( kms - ) log N (HI) Reference c ( h - kpc) - h Reference c (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)HE 0226 - . ± . ±
10 0 13 . ± .
06 (1) ... ...0.20701 0 14 . ± .
03 169 ± -
24 15 . ± .
04 (2) 32.0 - . + . ± .
05 (2) 92.3 - . - . . ± . ± < . > - . . ± . ± . ± . > - . . ± . ± . ± . > - . -
123 0.09180 +
20 13 . ± .
04 73 ± . ± .
04 (5) 80.2 - . - . - . . ± . ± . ± .
05 (5) 203.5 - . . ± .
07 361 ±
41 0 16 . ± .
05 (5),(7) 94.3 - . - . -
87 13 . ± . ± -
82 14 . ± .
05 (4),(5) ... > - . . ± . ±
15 0 14 . ± . . ± . ± . ± . > - . . ± . ±
16 0 14 . ± . - . +
069 0.12420 -
158 14 . ± . ± - > . - . -
82 14 . ± . - > . +
120 13 . ± . ± + > . +
188 14 . ± . +
188 14 . ± . . ± . ± . ± .
04 (4),(10) ... > - . a Galaxies are found at rp ≤ h - ∆ v ≤
300 kms - b In cases where no absorbing galaxies have been found, we place a conservative upper limit for MR based on the R -band threshold that corresponds to a 100% completeness out to ∆ θ < ′ in our spectroscopic survey. The magnitudethresholds are R = 23 for the field around HE0226 - R = 20 for the other two fields. c (1) Lehner et al. (2006); (2) Savage et al. (2005); (3) This work; (4) Thom & Chen (2008b); (5) Prochaska et al. (2004); (6) Prochaska et al. (2006); (7) Chen & Prochaska (2000); (8) Spinrad et al. (1993); (9) Howk et al. (2009); (10) Trippet al. (2008). absorbers with log N (O VI) ≥
14 have been identified withgalaxies at r p < h - kpc and ∆ v ≤
300 km s - , while onlytwo of the seven weaker O VI absorbers with log N (O VI) <
14 have been identified with galaxies at r p < h - kpc and ∆ v ≤
300 km s - . Given the small sample size, we cannotrule out possible correlations between absorber strengths andgalaxy properties or the possibility that some of the weakerO VI absorbers may originate in underdense IGM. For O VIabsorbers that do not have galaxies found at r p < h - kpc, we estimate the magnitude limit of the absorbing galax-ies based on our survey depth and present these numbers inColumn (9) of Table 8.8.2. The Origin of
O VI
Absorption Systems
The galaxy–O VI absorber cross-correlation analysis pre-sented in § 7.1 has clearly identified emission-line domi-nated galaxies to be principally responsible for the observedO VI absorbers, but the physical mechanism that distributesthe metal-enriched gas to large galactic distances is unclear.Two competing scenarios to explain the origin of O VI ab-sorbers in regions around emission-line galaxies are (1) star-burst driven outflows (e.g. Heckman et al. 2001; Kawata &Rauch 2007; Oppenheimer & Davé 2009) and (2) satellite ac-cretion (e.g. Wang 1993; Bournaud et al. 2004; Elmegreenet al. 2007). Given that these emission-line galaxies are pre-sumably low-mass (given the sub- L ∗ nature) and star-forming(given the presence of strong emission lines), the strong corre-lation seems qualitatively consistent with the expectations ofO VI originating in starburst outflows (c.f. Kawata & Rauch2007; Oppenheimer & Davé 2009). On the other hand, satel-lite accretions or galaxy interactions are also expected to in- duce star formation (see e.g. Li et al. 2008). Here we examinewhether additional insights for the origin of O VI absorberscan be gleaned based on known properties of the absorbinggalaxies.We have noted an apparent discrepancy between the ob-served partial covering fraction of O VI absorbing gas, κ ≈ r p . h - kpc andthe comparable clustering amplitudes of emission-line galax-ies and O VI absorbers on scales r p ≤ . h - Mpc. Whilethe partial covering fraction indicates that not all emission-line galaxies contribute to the observed O VI absorber statis-tics, the comparable clustering amplitude suggests otherwise.Namely, every emission-line galaxy has a corresponding O VIabsorber. Likewise, the comparable clustering amplitudes ofO VI and strong Ly α absorbers also imply that the two ab-sorber populations share common halos, but absorption-linesurveys show that not all strong Ly α absorbers are accompa-nied by an O VI absorber (see e.g. Tables 3&8). Althoughboth starburst outflows and tidal debris due to galaxy interac-tions may explain the observed partial covering fraction, it isnot clear that they can both explain the comparable clusteringamplitude at r p . h - Mpc.To understand these apparent discrepancies, we first notethat the correlation amplitude on scales r p < h - Mpc isdominated by galaxies that share common halos (e.g. Zhenget al. 2007; Tinker et al. 2007). The comparable clusteringamplitudes of O VI and emission-line galaxies (and strongLy α absorbers) found in our small sample from three QSOfields may be understood, if O VI absorbers arise preferen-tially in groups of emission-line galaxies. This hypothesis issupported by two features found in Figure 14. First, four oforrelation of Galaxies and QSO Absorption-line Systems 25the five galaxy–O VI absorber pairs are indeed found in anenvironment of multiple emission-line galaxies. Second, onlyone of the four “isolated” emission-line galaxies has a cor-reponding O VI absorber. Available observations are indeedconsistent with the expectations that O VI absorbers arise pri-marily in gas-rich galaxy groups. Taking into account themildly disturbed disk-like morphologies of O VI absorbinggalaxies, we further argue that the O VI absorbers originatein tidal debris produced by galaxy interactions in a group orpair environment and that tidal interactions may be principallyresponsible for distributing chemically enriched gas to largegalactic distances.A larger sample is necessary to improve the uncertaintiesin the measurements of κ for both group/pair and “isolated”galaxies, and to address possible bias due to field to field vari-ation. Extending the cross-correlation analysis to scales of r p ≈ h - Mpc is also necessary for constraining the meanhalo mass of the absorbers based on their large-scale cluster-ing amplitudes.8.3.
Extended Gas Around Galaxies
Previous studies have shown that luminous galaxies are sur-rounded by extended gaseous envelopes that may be probedby C IV and Mg II absorption transitions out to projectedphysical distances of ρ ≈ h - kpc (e.g. Steidel et al. 1994;Chen et al. 2001b; Tinker & Chen 2008; Chen & Tinker 2008)and by strong Ly α absorption out to ρ ≈ h - kpc (Lanzettaet al. 1995; Chen et al. 1998,2001a). The covering fractionof extended gas probed by the presence of these transitions isfound to be > r p ≈ h - kpc (corresponding to a projected physicaldistance ρ ≈ h - kpc at z = 0 .
3) with a mean covering frac-tion of κ ≈ α and Mg II absorption features arefound to be progressively weaker in galaxies at larger impactparameters (Chen et al. 1998, 2001a; Chen & Tinker 2008),we find a lack of such correlation in either C IV (Chen et al.2001b) or O VI (Figure 14). Including known O VI absorbinggalaxies from Cooksey et al. (2008) and Lehner et al. (2009)further increases the scatter in the W (1031) versus r p distribu-tion at r p < h - kpc.In a simple two-phase medium model, C IV and Mg II ab-sorbers arise in photo-ionized cold clumps pressure-confinedin a hot medium. The lack of correlation between C IV ab-sorber strength and galaxy impact parameter can thereforebe understood by higher gas pressure (and therefore reducedabundances of C + ) at smaller radii (e.g. Mo & Miralda-Escudé 1996). The same model cannot explain the presenceof O VI around galaxies of a broad range of luminosity dueto the high ionization potential. Instead, the mildly disturbeddisk morphologies observed in O VI absorbing galaxies sug-gest that tidal debris in small groups or close galaxy pairs maybe principally responsible for the observed O VI absorptionfeatures.Additional insights may be learned from comparisons withknown properties of the Milky Way halos. While exquisite de-tails, including tidal streams, infalling clouds, and outflows,have been recorded for extended gas around the Milky Way(e.g. Sembach et al. 2003), applications of these observationsfor constraining models of galaxy growth has been limited dueto a lack of knowledge for the distances to these Halo clouds.In contrast, observations of distant galaxy–absorbers pairs of- fer direct measurements of the distances between absorbinggas and star-forming regions, but the origin of the absorbingclouds (either due to accretion, outflow, or tidal stripping) isoften more uncertain.A systematic survey of O VI absorbers toward 100 extra-galactic sources has yielded positive detections in the Galacticthick disk and halo along 60 -
85% of these sightlines (Sem-bach et al. 2003). Because some of these detections are asso-ciated with known high velocity clouds at ≈
10 kpc distances(e.g. Thom et al. 2008), the observed incidence of 60 - ∼
250 kpc radius to an external observer.This partial covering of O VI gas in the halo of the MilkyWay, representing typical ∼ L ∗ galaxies in the nearby uni-verse, is qualitatively consistent with the partial covering frac-tion found in our study for halos around distant sub- L ∗ galax-ies. The lack of evidence for large-scale outflows in the MilkyWay halo based on detections of O VI absorbers also appearsto be consistent with our interpretation that the majority ofO VI absorbers are not produced in starburst outflows arounddistant star-forming galaxies. A larger sample of galaxy–O VIabsorber pairs that covers a broad range of impact parameterswill allow a conclusive characterization of the nature of theO VI gas, which in turn will shed light for the halo gas con-tent around the Milky Way.Finally, our analysis has also shown that the majority ofstrong Ly α and O VI absorbers do not probe the gaseous halosaround early-type galaxies. To study the warm-hot gas aroundearly-type galaxies or galaxy groups would require a differentprobe.
9. SUMMARY
We have carried out a deep, wide-area survey of galax-ies in fields around HE 0226 - - + α ab-sorbers of log N (H I) = 12 . - . N (O VI) = 13 . - . R -band magnitudes ofthese galaxies range from M R - h > -
16 to M R - h . -
22. The projected physical distances of these galaxies rangefrom ρ < h - kpc to ρ > h - Mpc. While our spec-troscopic survey has uncovered multiple large-scale galaxyoverdensities in the three fields, only two O VI absorbers (at z = 0 . -
123 and at z = 0 . + M R - h = - . - + ′ × ′ area roughly centered at theQSO. These images also allow us to examine the optical mor-phologies of individual absorbing galaxies.Combining various absorber and galaxy data, we have per-formed a cross-correlation study to understand the physicalorigin of Ly α and O VI absorbers and to constrain the proper-ties of extended gas around galaxies. The main results of our6 Chen & Mulchaeystudy are summarize below.1. Using a flux-limited sample of 670 foreground galax-ies within ∆ θ = 11 ′ of the QSO, we find based on across-correlation analysis that both strong Ly α absorbers oflog N (H I) ≥
14 and O VI absorbers exhibit a comparableclustering amplitude as emission-line dominated galaxies onscales of r p < h - co-moving Mpc. The clustering ampli-tudes of these absorbers are found to be ≈ α absorbers of log N (H I) < . α absorbers oflog N (H I) ≤
14 probe extended halos of emission-line domi-nated galaxies and that a large fraction of weak Ly α absorbersof log N (H I) ≤ . ≈ . L ∗ galaxies.2. O VI absorbers exhibit a similar behavior as strong Ly α absorbers. These absorbers also show a comparable cluster-ing amplitude as emission-line galaxies but a factor of sixlower clustering amplitude relative to absorption-line domi-nated galaxies on scales of r p ≤ h - Mpc. The results implythat the majority of O VI absorbers do not probe the gaseoushalos around massive, early-type galaxies, and that they probeprimarily halos around emission-line dominated star-forminggalaxies.3. Using a small sample of 20 R ≤
22 galaxies found atco-moving projected distances r p ≤ h - kpc from thelines of sight toward the three QSOs, we find that none ofthe absorption-line dominated galaxies in the sample has acorresponding O VI absorber to a sensitive upper limit of W (1031) . .
03 Å, and that the covering fraction of O VIabsorbing gas around emission-line dominated galaxies is κ ≈ r p < h - kpc and ∆ v ≤
300 km s - . On the other hand,two galaxy “groups” found at r p ≤ h - kpc from a QSOsightline do not have an associated O VI absorber. A moredetailed examination of galaxies in these latter two groupsshows that the most luminous members of both groups exhibitabsorption-line dominated spectral features, suggesting thatO VI may originate preferentially in gas-rich galaxy groups.5. Available high-resolution HST/WFPC2 images of fiveO VI absorbing galaxies show that these galaxies exhibit disk-like morphology with mildly disturbed features on the edge,suggesting that tidal disruption may be in effect.JSM would like to acknowledge the visitor’s program at theKavli Institute for Cosmological Physics, where part of thework presented here was completed. We thank helpful dis-cussions with J. Tinker and C. Thom. We thank L. Matthews,M. Rauch, A. Kravtsov for helpful comments on an early ver-sion of the manuscript. We thank R. Simcoe and A. Boltonfor assistance on obtaining part of the IMACS images pre-sented in this paper and R. Marzke for helpful discussions onreducing and assembling geometrically distorted multi-CCDimaging data. We are grateful to D. Kelson and G. Walth forassistance on the reduction of the IMACS spectra. We alsothank J. Scott for providing the combined FUSE spectrum ofPG 1216 + REFERENCESBertin, E. & Arnouts, S. 1996, A&AS, 117, 393Blanton, M. R. et al. 2003, ApJ 592, 819Bond, N. A., Churchill, C. W., Charlton, J. C., & Vogt, S. S. 2001, ApJ, 557,761Bournaud, F., Duc, P.-A., Amram, P., Combes, F., & Gach, J.-L. 2004, A&A,425, 813Carswell, R. F., Lanzetta, K. M., Parnell, H. C., & Webb, J. K. 1991, ApJ,371, 36Cen, R. & Ostriker, J. P. 2006, 650, 573Chen, H.-W., Lanzetta, K. M., Webb, J. K., & Barcons, X. 1998, ApJ, 498,77 . 2001a, ApJ, 559, 654Chen, H.-W. & Prochaska, J.X. 2000, ApJ, 543, L9Chen, H.-W., Lanzetta, K. M., & Webb, J. K. 2001b, ApJ, 556, 158Chen, H.-W., Prochaska, J. X., Weiner, B. J., Mulchaey, J. S., & Williger, G.M. 2005, ApJ, 629, L25Chen, H.-W. & Tinker, J. L. 2008, ApJ, 687, 745Churchill, C. W., Kacprzak, G. G., Steidel, C. C., & Evans, J. L. 2007, ApJ,661, 714Coleman, G. D., Wu, C. C., & Weedman, D. W. 1980, ApJS, 43, 393Cooksey, K. L., Prochaska, J. X., Chen, H.-W., Mulchaey, J. S., & Weiner, B.J. 2008, ApJ, 676, 262Davé, R., Cen, R., Ostriker, J. P., Bryan, G. L., Hernquist, L., Katz, N.,Weinberg, D. H., Norman, M. L., & O’Shea, B. 2001, ApJ, 552, 473Dobrzycki, A., Bechtold, J., Scott, J., & Morita, M. 2002, ApJ, 571, 654Dressler, A., Hare, T., Bigelow, B. C., & Osip, D. J. 2006, Proc. SPIE, 6269,13Elmegreen, D. M., Elmegreen, B. G., Ferguson, T., & Mullan, B. 2007, ApJ,663, 734Fukugita, M., Hogan, C.J, & Peebles, P.J.E. 1998, Nature, 503, 518Fukugita, M. 2004, in IAU Symp. 220, Dark Matter in Galaxies, ed. S. D.Ryder et al. (San Francisco: ASP), 227Grogin, N. A. & Geller, M. J. 1998, 505, 506Howk, J. C., Ribaudo, J. S., Lehner, N., Prochaska, J. X., & Chen, H.-W.2009, ApJ submittedJannuzi, B. T. et al. 1998, ApJS, 118, 1Kawata, D. & Rauch, M. 2007, ApJ, 663, 38Landolt, A.U. 1992, AJ, 104, 340 Landy, S. D. & Szalay, A. S. 1993, ApJ, 412, 64Lanzetta, K. M., Bowen, D. V., Tytler, D., & Webb, J. K. 1995, ApJ, 442, 538Lehner, N., Savage, B. D., Wakker, B. P., Sembach, K. R., & Tripp, T. M.2006, ApJS, 164, 1Lehner, N., Savage, B. D., Richter, P., Sembach, K. R., Tripp, T. M., &Wakker, B. P. 2007, ApJ, 658, 680Lehner, N., Prochaska, J. X., Kobulnicky, H. A., Cooksey, K. L., Howk, J. C.,Williger, G. M., & Cales, S. L. 2009, ApJ in press (arXiv:0812,4231)Li, C., Kauffmann, G., Heckman, T. M., White, S. D. M., & Jing, Y. P. 2008,MNRAS, 385, 1915Madgwick, D. S. et al. 2003, MNRAS, 344, 847Maller, A. H. & Bullock, J. S. 2004, MNRAS, 355, 694Mo, H. J. & Miralda-Escudé, J. 1996, ApJ, 469, 589Mo, H. J. & White, S. D. M. 2002, MNRAS, 336, 112Morris, S. L., Weymann, R. J., Dressler, A., McCarthy, P. J., Smith, B. A.,Terrile, R. J., Giovanelli, R., & Irwin, M. 1993, ApJ, 419, 524Mulchaey, J. S., Mushotzky, R. F., Burstein, D., & Davis, D. S. 1996, ApJ,456, L5Mulchaey, J. S. & Chen, H.-W. 2009, ApJ submittedOppenheimer, B. D. & Davé, R. 2009, MNRAS submitted (arXiv:0806.2866)Penton, S. V., Stocke, J. T., & Shull, J. M. 2002, ApJ, 565, 720. 2004, ApJS, 152, 29Persic, M. & Salucci, P. 1992, MNRAS, 258, 14Pickles, A. J. 1998, PASP, 110, 863Prochaska, J. X., Chen, H.-W., Howk, J. C., Weiner, B. J., & Mulchaey, J. S.2004, ApJ, 617, 718Prochaska, J. X., Weiner, B. J., Chen, H.-W., & Mulchaey, J. S. 2006, ApJ,642, 989Rauch, M., Miralda-Escudé, J., Sargent, W. L. W., Barlow, T. A., Weinberg,D. H., Hernquist, L., Katz, N., Cen, R., Ostriker, J. P. 1997, ApJ, 489, 7Rauch, M. 1998, ARA&A, 36, 267Savage, B. D., Lehner, N., Wakker, B. P., Sembach, K. R., & Tripp, T. M.2005, ApJ, 626, 776Scott, J. E., Kriss, G. A., Brotherton, M., Green, R. F., Hutchings, J., Shull, J.M., & Zheng, W. 2004, ApJ, 615, 135Sembach, K. R., Wakker, B. P., Savage, B. D., Richter, P., Meade, M., Shull,J. M., Jenkins, E. B., Sonneborn, G., & Moos, H. W. 2003, ApJS, 146, 165Sheth, R. K., Connolly, A. J., & Skibba, R. 2005, astro-ph/0511773 orrelation of Galaxies and QSO Absorption-line Systems 27orrelation of Galaxies and QSO Absorption-line Systems 27