Characterizing Circumgalactic Gas around Massive Ellipticals at z ~ 0.4 I. Initial Results
Hsiao-Wen Chen, Fakhri S. Zahedy, Sean D. Johnson, Rebecca M. Pierce, Yun-Hsin Huang, Benjamin J. Weiner, Jean-Rene Gauthier
MMNRAS , 1–19 (2017) Preprint 31 July 2018 Compiled using MNRAS L A TEX style file v3.0
Characterizing Circumgalactic Gas around MassiveEllipticals at z ∼ . I. Initial Results (cid:63)
Hsiao-Wen Chen † , Fakhri S. Zahedy , Sean D. Johnson , ‡ , Rebecca M. Pierce , ,Yun-Hsin Huang , Benjamin J. Weiner , and Jean-Ren´e Gauthier Department of Astronomy & Astrophysics, The University of Chicago, Chicago, IL 60637, USA Department of Astrophysics, Princeton University, Princeton, NJ 08544, USA The Observatories of the Carnegie Institution for Science, 813 Santa Barbara Street, Pasadena, CA 91101, USA Department of Aerospace Engineering, University of Maryland, College Park, MD 20742, USA Steward Observatory, University of Arizona, Tucson, AZ 85721, USA DataScience.com, Culver City, CA 90230, USA
31 July 2018
ABSTRACT
We present a new
Hubble Space Telescope ( HST ) Cosmic Origins Spectrograph(COS) absorption-line survey to study halo gas around 16 luminous red galaxies(LRGs) at z = 0 . − .
55. The LRGs are selected uniformly with stellar mass M star > M (cid:12) and no prior knowledge of the presence/absence of any absorptionfeatures. Based on observations of the full Lyman series, we obtain accurate mea-surements of neutral hydrogen column density N (H I ) and find that high- N (H I ) gasis common in these massive quiescent halos with a median of (cid:104) log N (H I ) (cid:105) = 16 . d < ∼
160 kpc. We measure a mean covering fraction of optically-thickgas with log N (H I ) > ∼ . (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
160 kpc and (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
100 kpc. The line-of-sight velocity separations between the H I absorbing gasand LRGs are characterized by a mean and dispersion of (cid:104) v gas − gal (cid:105) = 29 km s − and σ (cid:104) v gas − gal (cid:105) = 171 km s − . Combining COS far-ultraviolet and ground-based echellespectra provides an expanded spectral coverage for multiple ionic transitions, fromlow-ionization Mg II and Si II , to intermediate ionization Si III and C
III , and to high-ionization O VI absorption lines. We find that intermediate ions probed by C III and Si
III are the most prominent UV metal lines in LRG halos with a mean covering fractionof (cid:104) κ (C III ) (cid:105) . = 0 . +0 . − . for W r (977) (cid:62) . d <
160 kpc, comparable to whatis seen for C
III in L ∗ and sub- L ∗ star-forming and red galaxies but exceeding Mg II or O VI in quiescent halos. The COS-LRG survey shows that massive quiescent haloscontain widespread chemically-enriched cool gas and that little distinction betweenLRG and star-forming halos is found in their H I and C III content.
Key words: surveys – galaxies: haloes – galaxies: elliptical and lenticular, cD –quasars: absorption lines – intergalactic medium – galaxies: formation (cid:63)
Based on data gathered with the 6.5m Magellan Telescopeslocated at Las Campanas Observatory, the W. M. Keck Observa-tory, and the NASA/ESA Hubble Space Telescope operated bythe Space Telescope Science Institute and the Association of Uni-versities for Research in Astronomy, Inc., under NASA contractNAS 5-26555. † E-mail: [email protected] ‡ Hubble & Carnegie-Princeton Fellow
The circumgalactic medium (CGM), located in thespace between galaxies and the intergalactic medium (IGM),is regulated by the complex interaction between IGM accre-tion and stellar feedback. An accurate characterization ofthe CGM is therefore critical for understanding how galax-ies grow and evolve. Previous analytical calculations and hy-drodynamical simulations have suggested that chemically-pristine gas accreted from the IGM by low-mass galaxies c (cid:13) a r X i v : . [ a s t r o - ph . GA ] J u l Chen et al. in less massive dark matter halos ( M halo < ∼ M (cid:12) ) is cooland never shock-heated to the virial temperature of the darkmatter halo, whereas more massive halos acquire their gasthrough a hot channel in which the gas is shock-heated totemperatures T > ∼ K (see Faucher-Gigu`ere 2017 for arecent review and a list of references). At the same time,both cosmological simulations and observations of the high-redshift Universe have also highlighted the importance ofsupergalactic winds in enriching the CGM of star-forminggalaxies and the underdense IGM with heavy elements (seevan de Voort 2017 for a recent review and related references).The presence of heavy elements in turn alters the thermalstate of the gas.Over the last decade, substantial effort has been madeto understand the physical mechanisms driving the evolu-tion of the CGM (see Chen 2017a and Tumlinson et al. 2017for recent reviews), but the origin of the cool, T ∼ − K circumgalactic gas is largely uncertain. In particular, theobserved high incidence of chemically-enriched cool gas inthe vicinities of luminous red galaxies (LRGs; e.g., Gauthieret al. 2009, 2010, Lundgren et al. 2009; Gauthier & Chen2011; Bowen & Chelouche 2011; Zhu et al. 2014; Huang etal.2016) remains a puzzle, because LRGs are massive ellip-tical galaxies at z ∼ . > ∼ L ∗ and stellar masses of M star > ∼ M (cid:12) at z ≈ . M halo > ∼ M (cid:12) (e.g., Zheng et al. 2007; Blake etal. 2008; Padmanabhan et al. 2007). Independent studieshave also shown that more than 90% of massive galaxieswith M star > ∼ M (cid:12) in the local universe contain primarilyevolved stellar populations (e.g., Peng et al. 2010; Tinker etal. 2013), making the LRGs an ideal laboratory for studyingthe cold gas content in massive quiescent halos.Gauthier et al. (2009) noted that the cross-correlationsignal of photometrically selected LRGs and Mg II absorbers( z ∼ .
5) is comparable to the LRG auto-correlation onscales of r p < ∼
400 kpc. The comparable clustering amplitudeson scales of the LRG halo size suggests that a large fractionof the LRGs host a Mg II absorber, which is understood tooriginate in photo-ionized gas of temperature T ∼ K(Bergeron & Stas´ınska 1986; Charlton et al. 2003) and tracehigh-column density clouds of neutral hydrogen column den-sity N (H I ) ≈ − cm − (Rao et al. 2006). Subse-quent spectroscopic studies have indeed uncovered strongMg II absorbers in a significant fraction of LRG halos witha mean covering fraction of κ (Mg II ) > ∼
15 (5)% at projecteddistances d <
120 (500) kpc (Gauthier et al. 2010; Huanget al. 2016). In addition, strong Ly α absorption is foundcommon around early-type galaxies at (cid:104) z (cid:105) ≈ . z = 0 . − .
7, and found that high columndensity gas exists at projected distances as small as ≈ −
15 kpc from these ellipticals. While absorption-line studies con-tinue to show the presence of chemically-enriched cool gasaround red galaxies at intermediate redshifts, local 21 cmand CO surveys of elliptical galaxies have also uncoveredcold atomic and/or molecular gas in >
30% of nearby ellip-ticals (e.g., Oosterloo et al. 2010; Serra et al. 2012; Younget al. 2014). Together, these independent studies show thatcool gas outlasts star formation in massive quiescent halosover an extended cosmic time period.The presence of cool gas in massive halos challengestheoretical expectations of gaseous halos from both simpleanalytic models and numerical simulations. The predomi-nantly old stellar population and quiescent state in thesepassive galaxies also make it difficult to apply feedback dueto starburst and active galactic nuclei (AGN) as a generalexplanation for the presence of these absorbers (e.g., Rah-mati et al. 2015; Faucher-Gigu`ere et al. 2016). Applying thelarge number of SDSS Mg II absorption observations of LRGsat intermediate redshifts to test theoretical predictions re-quires knowledge of the ionization state and metallicity ofthe gas, which are not available in most cases. To facilitatedirect and quantitative comparisons between observationsand theoretical predictions, we are carrying out a systematicstudy of the CGM around LRGs using the Cosmic OriginsSpectrograph (COS) on board the Hubble Space Telescope ( HST ). The spectral coverage of COS enables observationsof the full H I Lyman series transitions and metal absorptionfeatures that probe halo gas under different ionization con-ditions, including C
III λ VI λλ , III λ II λ M = 0.3 and Ω Λ =0.7 with a Hub-ble constant H = 70 km s − Mpc − throughout the paper. To advance a deeper understanding of gas propertiesin massive quiescent halos, we are carrying out a COS-LRGprogram for a systematic study of the CGM around LRGs atintermediate redshifts. The primary goals of the program are(1) to obtain accurate and precise measurements of neutralhydrogen column density N (H I ) based on observations ofthe full hydrogen Lyman series and (2) to constrain the ion-ization state and chemical enrichment in massive quiescenthalos based on observations of a suite of ionic transitions.The LRGs in our sample are selected without prior knowl-edge of the presence/absence of any absorption features inthe halos and therefore enable an accurate assessment of thechemical enrichment level in these massive quiescent halos.Here we describe our program design and associated spec-troscopic data for the subsequent absorption-line studies. MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos Table 1.
Summary of the QSO–LRG Pair Sample
QSO LRGFUV θ d ID z QSO (mag) ID z LRG (arcsec) (kpc) ( u − g ) rest M r log M star /M a (cid:12) SDSSJ 094631.69+512339.9 0.741 18.4 SDSSJ 094632.40+512335.9 0.4076 7.7 41.7 1 . ± . − . ± .
06 11.2SDSSJ 140626.60+250921.0 0.867 18.3 SDSSJ 140625.97+250923.2 0.4004 8.8 47.3 1 . ± . − . ± .
04 11.1SDSSJ 111132.18+554726.1 0.766 17.5 SDSSJ 111132.33+554712.8 0.4629 13.2 77.1 1 . ± . − . ± .
07 11.4SDSSJ 080359.23+433258.3 0.449 18.4 SDSSJ 080357.74+433309.9 b . ± . − . ± .
02 11.1SDSSJ 092554.70+400414.1 0.471 18.2 SDSSJ 092554.18+400353.4 b . ± . − . ± .
02 11.1SDSSJ 095000.73+483129.2 0.590 17.9 SDSSJ 095000.86+483102.2 b . ± . − . ± .
02 11.0SDSSJ 112756.76+115427.1 0.509 18.1 SDSSJ 112755.83+115438.3 0.4237 17.7 98.7 1 . ± . − . ± .
06 11.2SDSSJ 124307.57+353907.1 0.547 18.4 SDSSJ 124307.36+353926.3 0.3896 19.3 102.2 1 . ± . − . ± .
04 11.3SDSSJ 155048.29+400144.9 0.497 18.1 SDSSJ 155047.70+400122.6 b . ± . − . ± .
02 11.2SDSSJ 024651.91 − − . ± . − . ± .
03 11.4SDSSJ 135726.27+043541.4 1.233 19.2 SDSSJ 135727.27+043603.3 0.3296 26.5 125.9 1 . ± . − . ± .
03 11.3SDSSJ 091029.75+101413.5 0.463 18.5 SDSSJ 091027.70+101357.2 b . ± . − . ± .
02 11.2SDSSJ 141309.14+092011.2 0.460 17.5 SDSSJ 141307.39+091956.7 0.3584 29.8 149.2 1 . ± . − . ± .
02 11.7SDSSJ 155304.92+354828.6 0.721 17.7 SDSSJ 155304.32+354853.9 0.4736 26.3 155.9 1 . ± . − . ± .
11 11.0SDSSJ 125901.67+413055.8 0.745 18.4 SDSSJ 125859.98+413128.2 0.2790 37.6 159.1 1 . ± . − . ± .
02 11.6SDSSJ 124410.82+172104.6 1.282 18.4 SDSSJ 124409.17+172111.9 0.5591 24.8 160.1 1 . ± . − . ± .
09 11.5 a Uncertainties in M star are known to be better than 0.2 dex (e.g., Conroy 2013). b COS-Halos red galaxies that match our LRG selection criteria; projected distances updated based on our own calculations.
To establish a uniform sample of LRGs for a comprehen-sive study of the ionization state and chemical enrichmentin massive quiescent halos, we cross-correlated UV-brightquasars with FUV < ∼ . z > ∼ . d < ∼
160 kpc fromthe sightline of a UV-bright background quasar to ensurethat high-resolution ( R ≡ λ/ ∆ λ ≈ Hubble SpaceTelescope ( HST ) for observing and resolving weak ionic ab-sorption features. The maximum projected distance d = 160kpc corresponds to roughly 1 / R vir ) ofa 10 M (cid:12) dark matter halo. It was motivated by the SDSSMg II survey of Huang et al. (2016), who found a mean cov-ering fraction of >
10% at d (cid:54)
160 kpc around LRGs for W r (2796) > . >
70% for weaker oneswith W r (2796) > . z > ∼ .
26 to ensure the ability to constrain neutral hy-drogen column density, N (H I ), from observations of higher-order Lyman absorption series. In addition, we restricted oursearch to LRGs that occur at a line-of-sight velocity differ-ence of ∆ v < − ,
000 km s − from the background QSO,to avoid confusion with outflowing gas from the quasar (e.g.,Wild et al. 2008).This exercise yielded a unique sample of 16 LRGs with aUV-bright background QSO found at d < ∼
160 kpc. Becauseboth UV bright QSOs and LRGs are rare, the number ofprojected LRG-QSO pairs with close separations is small.This modest sample size underscores the challenge of prob-ing the CGM in massive quiescent halos using absorption-line techniques. A summary of the LRG–QSO pair sample ispresented in Table 1, which lists for each pair the QSO ID,redshift ( z QSO ), and FUV magnitude observed by GALEXin the first three columns, and the LRG ID, redshift ( z LRG ),angular distance ( θ ) and the corresponding physical pro-jected distance ( d ) to the QSO sightline, the rest-frame u − g color, rest-frame r -band magnitude ( M r ), and total stellarmass ( M star ) in the following seven columns. For each LRG, the rest-frame u − g color and M r were estimated by inter-polating between available optical broad-band photometryin the observed SDSS u , g , r , i , and z bandpasses, and M star was estimated using the K -correct code (Blanton & Roweis2007) for a Chabrier Initial Mass Function (Chabrier 2003).Error bars in u − g and M r reflect 1- σ uncertainties in thephotometric measurements. Uncertainties in M star are un-derstood to be dominated by systematic differences in themodel priors and are known to be less than 0.2 dex (e.g.,Moustakas et al. 2013; Conroy 2013).Five of the LRGs in our sample overlap with the COS-Halos red galaxy subsample (e.g., Tumlinson et al. 2011;Thom et al. 2012). Comparisons of our COS-LRG samplewith the red galaxies studied by the COS-Halos team (e.g.,Thom et al. 2012) are presented in Figure 1, which dis-plays the M star versus d distribution in the top panel andthe M star versus z distribution in the bottom panel. Notethat we have re-calculated M star for all COS-Halos galaxiesbased on available SDSS photometric measurements. Ourestimates of M star using K -correct are found to be system-atically lower than those reported in Werk et al. (2012)by 0.2 dex. The discrepancy can be attributed to a differ-ent stellar initial mass functions adopted by these authors.For consistency, we adopt our M star estimates for all COS-Halos galaxies in comparisons with the COS-LRG sample.In addition, one galaxy in the COS-Halos red galaxy sam-ple (J1617+0638 253 39) satisfies our mass selected criterionbut is not included in the COS-LRG sample. This red galaxyoccurs at d = 101 kpc from the background QSO sightlineand has M star ≈ × M (cid:12) , but the redshift is too lowthat the Lyman series transitions are not covered by avail-able COS spectra. While Ly α is not detected with a 2- σ upper limit to the rest-frame absorption equivalent width of W r (1215) = 0 .
14 ˚A (Thom et al. 2012), we exclude this fromthe COS-LRG in order not to bias subsequent analysis onthe incidence of cool gas in these massive halos. We also notethat one of the COS-Halos red galaxies, J1133+032 110 5 at z = 0 .
237 and d = 18 kpc (red diamond with a dotted circlein Figure 1), has a neighboring QSO at the same redshiftand d = 124 kpc. Quasar host halos have been shown to ex-hibit CGM properties that are different from both starburst MNRAS000
237 and d = 18 kpc (red diamond with a dotted circlein Figure 1), has a neighboring QSO at the same redshiftand d = 124 kpc. Quasar host halos have been shown to ex-hibit CGM properties that are different from both starburst MNRAS000 , 1–19 (2017)
Chen et al.
COS-Halos RedCOS-LRG
Figure 1.
Comparisons of the LRGs in our studies (orange filledcircles) and the red galaxies in the COS-Halos sample (red di-amonds) considered by Thom et al. (2012). The LRGs forma unique, mass-limited sample of quiescent galaxies located at d < ∼
160 kpc from a UV bright background QSO. M star of theLRGs ranges from log M star = 11 to log M star / M (cid:12) = 11 . (cid:104) M star / M (cid:12) (cid:105) COS − LRG = 11 .
2. In contrast,the COS-Halos red galaxies span a wide range in M star fromlog M star / M (cid:12) = 10 . M star / M (cid:12) = 11 . (cid:104) M star / M (cid:12) (cid:105) COS − HalosRed = 10 .
8. Note that one of the COS-Halos red galaxies, J1133+032 110 5 at z = 0 .
237 and d = 18 kpc(red diamond with a dotted circle), has a neighboring QSO at thesame redshift and d = 124 kpc. Five of the remaining COS-Halosred galaxies satisfy the LRG selection criteria, and are thereforeincluded in the LRG sample. These are shown in both sets ofsymbols. We have re-calculated both M star and d ourselves forconsistency, which led to small offsets in d . and passive galaxies (e.g., Prochaska et al. 2014; Johnson etal. 2015a; see also Chen 2017b for relevant discussions).Figure 1 demonstrates that the LRGs form a uni-form sample of high-mass halos with M star ranging fromlog M star / M (cid:12) = 11 to log M star / M (cid:12) = 11 . (cid:104) log M star / M (cid:12) (cid:105) COS − LRG = 11 . z = 0 .
21 to z = 0 .
55. In con-trast, the COS-Halos red galaxy sample includes pre-dominantly lower-mass halos with M star ranging fromlog M star / M (cid:12) = 10 . M star / M (cid:12) = 11 . (cid:104) log M star / M (cid:12) (cid:105) COS − HalosRed = 10 . z = 0 .
14 to z = 0 .
27. The COS-LRG sampletherefore offers a unique opportunity both for studying the
Table 2.
Journal of
HST
COS Observations
QSO Grating t exp (s) PIDSDSSJ 024651.91 − CGM in the most massive individual galaxy halos and forcomparison with less massive but still quiescent systems.
High-resolution FUV absorption spectra of the back-ground QSOs are available either from our own COS pro-gram (PID = 14145) or in the
HST
COS data archive. COSwith the G130M and G160M gratings and different cen-tral wavelengths offers a contiguous spectral coverage from λ ≈ λ ≈ ≈
17 km s − . Thespectral coverage of COS enables observations of the fullH I Lyman series transitions and metal absorption featuresthat probe halo gas under different ionization conditions.Common metal absorption lines observable by COS includeC III λ VI λλ , III λ II λ HST archive for SDSSJ 124410.82+172104.6,the LRG in this field occurs at sufficiently high redshift( z LRG = 0 . CALCOS pipeline were retrieved from the
HST archive, and further processed using our own software. Thewavelength calibration for COS spectra has been shown tocontain wavelength-dependent errors up to ≈
15 km s − (e.g., Johnson et al. 2013; Liang & Chen 2014; Wakker etal. 2015), leading to apparent misalignments of spectral fea-tures between different exposures. Additional wavelengthcalibration effort is therefore needed both to optimize thesignal-to-noise ( S/N ) in the final combined spectrum andto ensure the accuracy in measuring the kinematic proper-ties of different absorption features.We developed our own software to perform this neces-sary wavelength calibration in two steps. First, we correctedthe relative wavelength offsets between individual spectrausing a low-order polynomial that best describes the off-sets between common absorption lines found in different
MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos exposures. The common absorption lines include both in-tervening absorbers at z > III λ II λ II λ II λ absolute wave-length correction was performed using a low-order polyno-mial by registering unsaturated low-ionization Milky Wayfeatures to their known wavelengths in vacuum. In gen-eral, the absolute wavelength calibration for the G160Mspectra was more challenging than the G130M data due tothe fact that fewer unsaturated lines were available. Thefinal solution was evaluated and verified using available Ly-man series of strong intervening H I absorbers across thefull spectral range. In some cases, a new solution for theG160M spectrum was obtained using only a constant wave-length offset that was determined based on absorption fea-tures in the overlapping window between the G130M andG160M data. When comparing the velocity centroids be-tween low-ionization lines observed in COS and those ofMg II λλ , ≈
10 km s − ; see the discussion in § ± − . Finally, eachcombined spectrum was continuum normalized using a low-order polynomial fit to spectral regions free of strong ab-sorption features. The final continuum-normalized spectrahave a median S/N ≈ −
30 per resolution element.
Optical echelle spectra of the background QSOsare available for 11 of the 16 QSOs in the COS-LRGsample. These echelle spectra extend the wavelengthcoverage of absorption spectroscopy from λ > ∼ λ > ∼ II absorption series, theMg II λλ , I λ − (cid:48)(cid:48) slit and 2 × ≈
11 km s − at wavelength λ < λ ≈ λ ≈ Table 3.
Journal of Optical Echelle Observations
QSO Instrument t exp (s) S/N
SDSSJ 024651.91 − Optical echelle spectra of SDSSJ 080359.23+433258.3,SDSSJ 091029.75+101413.5, SDSSJ 092554.70+400414.1,SDSSJ 095000.73+483129.2, SDSSJ 155048.29+400144.9,and SDSSJ 155304.92+354828.6 obtained using HIRES(Vogt et al. 1994) on the Keck Telescopes are available inthe Keck Observatory Archive (KOA). The spectra wereobtained using a 0 . (cid:48)(cid:48) slit and 2 × ≈ . − atwavelength λ < λ ≈ λ ≈ S/N perresolution element at λ <
The continuum-normalized UV and optical QSO spec-tra described in § ±
500 km s − from the systemicredshift of the LRG. This search window corresponds to ≈ ± σ v , where σ v is the observed line-of-sight velocity dis-persion of Mg II absorbing gas in LRG halos (e.g., Huang etal. 2016). It is therefore sufficiently large to include absorp-tion features physically connected to the LRGs.To characterize the gas content in LRG halos, we ob-tain two sets of measurements. First, we perform a Voigtprofile analysis, taking into account the presence of multipleabsorption components per halo, and determine N (H I ) andthe Doppler parameter b H I of individual components, andthe associated measurement uncertainties. Next, we measurethe total, integrated absorption equivalent widths of differ-ent ionic absorption transitions, which enable direct com-parisons between absorption properties of the LRG halos inour study and those of quiescent halos studied previously(e.g. Gauthier & Chen 2011; Thom et al. 2012; Werk et al.2013; Huang et al. 2016). Here we describe the absorption MNRAS000
500 km s − from the systemicredshift of the LRG. This search window corresponds to ≈ ± σ v , where σ v is the observed line-of-sight velocity dis-persion of Mg II absorbing gas in LRG halos (e.g., Huang etal. 2016). It is therefore sufficiently large to include absorp-tion features physically connected to the LRGs.To characterize the gas content in LRG halos, we ob-tain two sets of measurements. First, we perform a Voigtprofile analysis, taking into account the presence of multipleabsorption components per halo, and determine N (H I ) andthe Doppler parameter b H I of individual components, andthe associated measurement uncertainties. Next, we measurethe total, integrated absorption equivalent widths of differ-ent ionic absorption transitions, which enable direct com-parisons between absorption properties of the LRG halos inour study and those of quiescent halos studied previously(e.g. Gauthier & Chen 2011; Thom et al. 2012; Werk et al.2013; Huang et al. 2016). Here we describe the absorption MNRAS000 , 1–19 (2017)
Chen et al. line analysis for measuring N (H I ) and for constraining theincidence of heavy ions in massive quiescent halos. N ( H I ) and b HI The COS-LRG sample is established based on the crite-rion that robust constraints for N (H I ) can be placed basedon observations of the Lyman series lines. Specifically, obser-vations of the flux decrement at the Lyman limit transitionprovide a direct constraint for N (H I ), while observations ofthe Lyman absorption lines constrain the Doppler parameter b H I . In Figures 2a&b, we present the continuum-normalizedfinal spectra of the LRG halos over the rest-frame spectralrange from λ < ∼
910 ˚A to Ly γ , organized with increasing pro-jected distance from top to bottom.For all but two sightlines (SDSSJ 080359.23+433258.3and SDSSJ 092554.70+400414.1), the available COS dataprovide a complete spectral coverage for the Lyman se-ries transitions from Ly α to below Lyman limit . ForSDSSJ 080359.23+433258.3, the LRG is at z = 0 . λ ≈
930 ˚A at the rest frame of the LRG. How-ever, because the absorbing gas along the QSO sightlinein this LRG halo is optically thin, observations of Ly α ,Ly β , Ly γ and Ly δ are sufficient for constraining the N (H I )without additional higher-order Lyman series lines. ForSDSSJ 092554.70+400414.1, the LRG occurs at z = 0 . α absorption lineexhibits prominent damping wings, an accurate measure-ment of N (H I ) can also be obtained based on the dampedprofiles.For each LRG halo, we perform a Voigt profile anal-ysis using a custom software. We first generate a modelabsorption spectrum, including the full Lyman series andcontinuum absorption beyond the Lyman limit, based onthe minimum number of discrete components needed toexplain the observed absorption profiles. Each componentis uniquely defined by three parameters: N (H I ), b H I , andthe velocity centroid dv c relative to the redshift of thestrongest absorbing component z abs . The theoretical Lymanseries spectrum is then convolved with the COS line spreadfunction, appropriate for the lifetime position the spectrawere obtained at, and binned to the adopted pixel resolu-tion of the data. Next, the model spectrum is compared tothe continuum-normalized COS spectrum, and the best-fitmodel is found using a χ minimization routine. Finally,we perform a Markov Chain Monte Carlo (MCMC) analy-sis using the emcee package (Foreman-Mackey et al. 2013),in order to assess the probability distribution of the best-fit model parameters. The MCMC run consists of 200 stepswith an ensemble of 250 walkers, which are initialized overa small region in the parameter space around the minimum χ value. The MCMC analysis allows us to construct the The LRG toward SDSSJ 124409.17+172111.9 is found at z =0 . α transition occurs at λ = 1895˚A. Therefore, only Ly β and higher Lyman series lines are observedin the COS FUV channel. marginalized probability distribution for each parameter ofthe model. The sum of the best-fit Voigt profiles of indi-vidual components is displayed in red solid line for eachLRG halo in Figures 2a&b. The best-fit N (H I ) and b H I of each component and the corresponding 68% confidenceintervals are presented in columns (8) and (9) of Table 4,along with z abs , relative velocity v gas − gal of z abs to z LRG , dv c , and the total N (H I ) summed over all components incolumns (4) through (8). Note that while the search win-dow in the line-of-sight velocity is large, ±
500 km s − , allthe identified absorption features fall well within this searchwindow with a mean and dispersion between the strongestH I component and the LRG of (cid:104) v gas − gal (cid:105) = 29 km s − and σ (cid:104) v gas − gal (cid:105) = 171 km s − , consistent with the velocity dis-persion found between Mg II absorbing gas and LRGs in thelarger SDSS LRG sample in Zhu et al. (2014) and in Huanget al. (2016).Three of the 16 LRGs do not exhibit any correspondingLyman absorption series in the search window. These areSDSSJ 124307.36+353926.3 at z = 0 . d = 102 kpc,SDSSJ 141307.39+091956.7 at z = 0 . d = 149 kpc,and SDSSJ 125859.98+413128.2 at z = 0 . d = 159kpc. We place a sensitive 2- σ upper limit of N (H I ) < . ±
16 km s − around the ex-pected Ly α location, approximately twice the spectral reso-lution element of the COS spectra, using the associated 1- σ error spectrum for these LRG halos.Our analysis shows that high- N (H I ) gas is common inLRG halos with a median of (cid:104) log N (H I ) (cid:105) med = 16 . d < ∼
160 kpc. The spectral coverage of the rest-frame Ly-man limit transition proves to be critical for constraining N (H I ) in these Lyman limit and partial Lyman limit sys-tems. The excellent agreement between the observed andbest-fit absorption profiles of the Lyman series lines and theflux discontinuity at the Lyman limit in Figure 2a&b demon-strates that our best estimated N (H I ) and b H I in Table 4 areindeed accurate. In one case, SDSSJ 155048.29+400144.9 at z = 0 . d = 107 kpc, we note that the best-fit N (H I )successfully reproduces the full Lyman series lines but pre-dicts a larger flux decrement below the Lyman limit thanwhat is observed in the data by more than 1- σ . Inspectingthe acquisition image of the COS observations, we find a sec-ond point source in the COS aperture at ≈ . (cid:48)(cid:48) along thedispersion direction from the QSO. The excess flux belowthe Lyman limit of the LRG can be explained if the red-shift of this second source is below the redshift of the LRG.We also note a second case, SDSSJ 124409.17+172111.9 at z = 0 . d = 160 kpc, for which the continuumflux declines to zero below 909 ˚A in the rest frame ofthe LRG. A Mg II absorber of rest-frame equivalent width W r (2796) ≈ . z = 0 . II absorber containsa large neutral gas column of log N (H I ) = 19 . ± . z = 0 . MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos kpc d = 42 47 77 79 84 94 99 102 h i g h - o r d e r L y m a n s e r i e s H I H I F i g u r e a . T h e L y m a n s e r i e s a b s o r p t i o n s p ec t r ao b s e r v e d a t d i ff e r e n t p r o j ec t e dd i s t a n ce s f r o m L R G s , f r o m d = k p c a tt h e t o p a nd i n c r e a s i n g t o d = k p c a tt h e b o tt o m . T h ec o n t i nuu m - n o r m a li ze d s p ec t r aa r e s h o w n i nb l a c k w i t h t h ec o rr e s p o nd i n g1 - σ e rr o r s h o w n i n c y a n . T h e g r ee n a nd r e dd a s h - d o tt e d li n e s m a r k t h e n o r m a li ze d c o n t i nuu m a nd ze r o flu x l e v e l s f o r g u i d a n ce . F o r e a c h L R G h a l o , t h e v e l o c i t y p r o fi l e s o f L y γ , a nd L y δ a r e p r e s e n t e d i n t h e t w o r i g h t p a n e l s w i t h ze r o v e l o c i t y c o rr e s p o nd i n g t o z a b s i n T a b l e , a nd t h e r e m a i n i n g h i g h e r - o r d e r L y m a n s e r i e s li n e s , a l o n g w i t h t h e L y m a n li m i t , a r e p r e s e n t e d i n t h e l e f t p a n e l w i t h t h e b l u e d o tt e d li n e s i nd i c a t i n g t h ee x p ec t e dp o s i t i o n s o f t h e L y m a n t r a n s i t i o n s . Sp ec t r a l r e g i o n s t h a t a r e n o t c o v e r e db y a v a il a b l e C O S s p ec t r aa r e g r e y e d o u t . T h e b e s t - fi t L y m a n s e r i e s a b s o r p t i o n s p ec t r aa r e s h o w n i n r e d MNRAS000
GM in Massive Quiescent Halos kpc d = 42 47 77 79 84 94 99 102 h i g h - o r d e r L y m a n s e r i e s H I H I F i g u r e a . T h e L y m a n s e r i e s a b s o r p t i o n s p ec t r ao b s e r v e d a t d i ff e r e n t p r o j ec t e dd i s t a n ce s f r o m L R G s , f r o m d = k p c a tt h e t o p a nd i n c r e a s i n g t o d = k p c a tt h e b o tt o m . T h ec o n t i nuu m - n o r m a li ze d s p ec t r aa r e s h o w n i nb l a c k w i t h t h ec o rr e s p o nd i n g1 - σ e rr o r s h o w n i n c y a n . T h e g r ee n a nd r e dd a s h - d o tt e d li n e s m a r k t h e n o r m a li ze d c o n t i nuu m a nd ze r o flu x l e v e l s f o r g u i d a n ce . F o r e a c h L R G h a l o , t h e v e l o c i t y p r o fi l e s o f L y γ , a nd L y δ a r e p r e s e n t e d i n t h e t w o r i g h t p a n e l s w i t h ze r o v e l o c i t y c o rr e s p o nd i n g t o z a b s i n T a b l e , a nd t h e r e m a i n i n g h i g h e r - o r d e r L y m a n s e r i e s li n e s , a l o n g w i t h t h e L y m a n li m i t , a r e p r e s e n t e d i n t h e l e f t p a n e l w i t h t h e b l u e d o tt e d li n e s i nd i c a t i n g t h ee x p ec t e dp o s i t i o n s o f t h e L y m a n t r a n s i t i o n s . Sp ec t r a l r e g i o n s t h a t a r e n o t c o v e r e db y a v a il a b l e C O S s p ec t r aa r e g r e y e d o u t . T h e b e s t - fi t L y m a n s e r i e s a b s o r p t i o n s p ec t r aa r e s h o w n i n r e d MNRAS000 , 1–19 (2017)
Chen et al. kpc d = 107 109 126 140 149 156 159 160 h i g h - o r d e r L y m a n s e r i e s H I H I F i g u r e b . S a m e a s F i g u r e , bu t f o r L R G h a l o s a t p r o j ec t e dd i s t a n ce s f r o m d = k p c a tt h e t o p a nd i n c r e a s i n g t o d = k p c a tt h e b o tt o m . N o t e t h a t f o r S D SS J . + . i n t h e t o pp a n e l t h e b e s t - fi t N ( H I ) s u cce ss f u ll y r e p r o du ce s t h e f u ll L y m a n s e r i e s li n e s bu t p r e d i c t s a l a r g e r flu x d ec r e m e n t b e l o w t h e L y m a n li m i t t h a n w h a t i s o b s e r v e d i n t h e d a t a b y m o r e t h a n - σ . T h e a c q u i s i t i o n i m ag e o f t h e C O S o b s e r v a t i o n s r e v e a l s a s ec o ndp o i n t s o u r ce i n t h e C O S a p e r t u r e a t ≈ . (cid:48)(cid:48) a l o n g t h e d i s p e r s i o nd i r ec t i o n f r o m t h e Q S O , w h i c h m a y h a v ec o n t r i bu t e d t o t h ee x ce ss flu x b e l o w t h e L y m a n li m i t o f t h e L R G . MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos In addition to the hydrogen Lyman series absorptionfeatures, we also search for associated ionic transitions forconstraining the incidence of heavy ions in massive quiescenthalos, including Mg II λλ , II λ II λ III λ III λ VI λλ , II doublet,Si II , Si III , C
III , and the O VI doublet transitions found ineach LRG halo, along with the hydrogen Ly α and Ly β ab-sorption lines, both observations and best-fit Voigt profilemodels, at the top of each column for comparison. For thethree LRGs with no H I absorption detected, we includedthe Voigt profiles for absorbing gas of log N (H I ) = 12 . b H I = 15 km s − for comparison. Zero velocity corre-sponds to z abs in Table 4. For clarity, spectral regions thatare not covered by available COS spectra or contain con-taminating features are greyed out. For H I , Mg II , Si II , andO VI lines, two transitions are observed which enable iden-tifications of contaminating features based on their knownrelative line strengths. For Si III and C
III , we rely on thematching velocity profiles with other low-ionization lines toidentify contaminating features.We note the excellent agreement in the velocity cen-troids between individual components of different ionic fea-tures, such as the Mg II absorption doublets and Si II , Si III ,and C
III . Recall the issues with COS wavelength calibra-tion described in § II absorption lines,which were calibrated using a combination of comparisonlamp and sky lines. The observed excellent agreement be-tween Mg II absorption lines and COS observations of Si II ,Si III , and C
III lines therefore indicates that the relative mo-tion between the Milky Way gas and the LSR toward thesesightlines is indeed negligible. In addition, the matching ve-locity profiles between H I , Mg II , and the remaining FUVlines also provide an important additional guide for filteringout contaminating features.Two features are immediately clear from Figures 3a&b.First of all, heavy ions are common in these massive qui-escent halos. In all but one LRG with Ly α and Ly β de-tected, metal absorption features are also detected (see § v ≈
200 km s − in three cases, e.g.,SDSSJ 094632.40+512335.9 at z = 0 . d = 42 kpc,SDSSJ 024651.20 − z = 0 . d = 109kpc, and SDSSJ 091029.75+101413.5 at z = 0 . d = 140 kpc, the line profiles of individual componentsfor all LRG halos are narrow. The observed narrow com-ponent profiles suggest that the gas is photo-ionized. Theonly exception is the O VI doublet transitions detected inthe halo of SDSSJ 094632.40+512335.9, which appear broad(FWHM ≈
300 km s − ) and asymmetric, indicating a dif-ferent physical origin from the associated hydrogen and low-ionization gas. We measure the rest-frame total, integrated absorptionequivalent widths of both Ly α and associated ionic lines, in-cluding Mg II λ II λ III λ III λ VI λ σ upper limitto the underlying absorption feature over a wavelength win-dow of ±
16 km s − around the expected line location, usingthe associated error spectrum. The wavelength window ischosen to cover approximately twice the spectral resolutionelement of the COS spectra around the expected lines. Theabsorption equivalent width measurements and associateduncertainties are presented in columns (9) through (14) ofTable 4, along with 2- σ upper limits for non-detections. En-tries with “...” indicate that either the absorption transitionsare not covered by available data or they are blended withcontaminating features and therefore no equivalent widthmeasurements are possible. For the five LRGs that have alsobeen studied by the COS-Halos team, equivalent width mea-surements of these transitions have also been published inWerk et al. (2013). Good agreements, to within 2- σ mea-surement uncertainties are found in all but one transition.For SDSSJ 091029.75+101413.5 at z = 0 . d = 140kpc, Werk et al. (2013) reported W r = 282 ±
63 m˚A forC
III λ W r (977) = 681 ±
48 m˚A. Com-paring the continuum-normalized absorption profiles of ourown and those presented in Figure 3 of Werk et al. (2013),we find that the discrepancy is likely due to a combination oflow
S/N and uncertainties in the continuum normalization.
The absorption-line search described in § I absorbers in a large fraction of LRG ha-los at z = 0 . − .
55. Observations of the full hydrogenLyman series have enabled accurate and precise measure-ments of N (H I ). The median H I column density is foundto be (cid:104) log N (H I ) (cid:105) = 16 . d < ∼
160 kpc (or d < ∼ / R vir )in these LRG halos. The line-of-sight velocity separationsbetween the H I absorbing gas and LRGs are characterizedby a mean and dispersion of (cid:104) v gas − gal (cid:105) = 29 km s − and σ (cid:104) v gas − gal (cid:105) = 171 km s − . The observed velocity dispersionof H I gas is similar to what is observed for Mg II absorbinggas (e.g., Zhu et al. 2014; Huang et al. 2016), but less thanthe expected line-of-sight velocity dispersion for virializedgas in these massive haloes. A suppressed velocity disper-sion implies that the kinetic energy of the absorbing clumpsis being dissipated, possibly due to the ram-pressure dragforce as the clumps move through the hot halo. We refer thereaders to Gauthier & Chen (2011) and Huang et al. (2016)for a discussion on the implied mass limit for the clumps.In addition to accurate N (H I ) measurements, we havealso been able to obtain accurate constraints for b H I ofindividual components. The best-estimated b H I values foroptically-thick components (log N (H I ) > .
2) are all rel-atively narrow with b H I < ∼
20 km s − , indicating a rela-tively cool gas temperature of T < ∼ × K. Combin-
MNRAS , 1–19 (2017) Chen et al. ing the observed H I absorption profiles with those of low-,intermediate-, and high-ionization species provides a uniqueopportunity for constraining the ionization condition, ther-mal state, and chemical abundances of the CGM in massivequiescent halos. A detailed ionization analysis is presentedin a subsequent paper (Zahedy et al. 2018, in preparation).Here we discuss general halo gas properties observed aroundLRGs. MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos d = J k p c J J J J J J J Normalized Flux H I H I M g II M g II S i II S i II S i III C III O V I O V I R e l a t i v e V e l o c i t y ( k m / s ) F i g u r e a . A b s o r p t i o np r o fi l e s o f ( f r o m t o p t o b o tt o m ) L y α , L y β , a nd a ss o c i a t e d M g II , S i II , S i III , C III , a nd O V I f o r e i g h t L R G h a l o s a t d < k p c ( i n c r e a s i n g d f r o m l e f tt o r i g h t) . F o ll o w i n g F i g u r e , ze r o v e l o c i t y c o rr e s p o nd i n g t o z a b s i n T a b l e , - σ e rr o r s a r e s h o w n i n c y a n . Sp ec t r a l r e g i o n s t h a t a r ee i t h e r c o n t a m i n a t e d o r n o t c o v e r e db y a v a il a b l e a b s o r p t i o n s p ec t r aa r e g r e y e d o u t f o r c l a r i t y . S y s t e m i c v e l o c i t i e s o f t h e L R G s a r e m a r k e db y t h e v e r t i c a l d o tt e d li n e , a nd t h e s p ec t r a l w i nd o w s u s e d f o r e q u i v a l e n t w i d t h m e a s u r e m e n t s i n T a b l e r e m a r k e d w i t h a h o r i z o n t a l b a r a tt h e t o p o f e a c hp a n e l. F o r t h e L R G a t d = k p c , n oa b s o r p t i o n f e a t u r e s a r e d e t ec t e d a nd - σ upp e r li m i t s a r e o b t a i n e d w i t h i n ± s p ec t r a l r e s o l u t i o n e l e m e n t , t h e w i nd o w m a r k e db y t h e t h i c k h o r i z o n t a l b a r . F o r t h e b r oa d d a m p e d L y α f e a t u r e a t d = k p c , W r ( ) w a s m e a s u r e d f r o m − t o600 k m s − . T h e b e s t - fi t L y α a nd L y β V o i g t p r o fi l e s , b o t h f o r i nd i v i du a l c o m p o n e n t s s e p a r a t e l y (t h i n li n e s ) a nd f o r a ll c o m p o n e n t s c o m b i n e d (t h i c k r e d li n e ) , o f e a c h L R G h a l oa r e a l s o r e p r o du ce d i n t h e t o p t w o p a n e l s f o r d i r ec t c o m p a r i s o n s w i t h r e s o l v e d m e t a l - li n ec o m p o n e n t s . A ll L R G s a t d < k p ce x h i b i t b o t hh y d r og e n a nd a ss o c i a t e d m e t a l - li n e a b s o r p t i o n f e a t u r e s . MNRAS000
GM in Massive Quiescent Halos d = J k p c J J J J J J J Normalized Flux H I H I M g II M g II S i II S i II S i III C III O V I O V I R e l a t i v e V e l o c i t y ( k m / s ) F i g u r e a . A b s o r p t i o np r o fi l e s o f ( f r o m t o p t o b o tt o m ) L y α , L y β , a nd a ss o c i a t e d M g II , S i II , S i III , C III , a nd O V I f o r e i g h t L R G h a l o s a t d < k p c ( i n c r e a s i n g d f r o m l e f tt o r i g h t) . F o ll o w i n g F i g u r e , ze r o v e l o c i t y c o rr e s p o nd i n g t o z a b s i n T a b l e , - σ e rr o r s a r e s h o w n i n c y a n . Sp ec t r a l r e g i o n s t h a t a r ee i t h e r c o n t a m i n a t e d o r n o t c o v e r e db y a v a il a b l e a b s o r p t i o n s p ec t r aa r e g r e y e d o u t f o r c l a r i t y . S y s t e m i c v e l o c i t i e s o f t h e L R G s a r e m a r k e db y t h e v e r t i c a l d o tt e d li n e , a nd t h e s p ec t r a l w i nd o w s u s e d f o r e q u i v a l e n t w i d t h m e a s u r e m e n t s i n T a b l e r e m a r k e d w i t h a h o r i z o n t a l b a r a tt h e t o p o f e a c hp a n e l. F o r t h e L R G a t d = k p c , n oa b s o r p t i o n f e a t u r e s a r e d e t ec t e d a nd - σ upp e r li m i t s a r e o b t a i n e d w i t h i n ± s p ec t r a l r e s o l u t i o n e l e m e n t , t h e w i nd o w m a r k e db y t h e t h i c k h o r i z o n t a l b a r . F o r t h e b r oa d d a m p e d L y α f e a t u r e a t d = k p c , W r ( ) w a s m e a s u r e d f r o m − t o600 k m s − . T h e b e s t - fi t L y α a nd L y β V o i g t p r o fi l e s , b o t h f o r i nd i v i du a l c o m p o n e n t s s e p a r a t e l y (t h i n li n e s ) a nd f o r a ll c o m p o n e n t s c o m b i n e d (t h i c k r e d li n e ) , o f e a c h L R G h a l oa r e a l s o r e p r o du ce d i n t h e t o p t w o p a n e l s f o r d i r ec t c o m p a r i s o n s w i t h r e s o l v e d m e t a l - li n ec o m p o n e n t s . A ll L R G s a t d < k p ce x h i b i t b o t hh y d r og e n a nd a ss o c i a t e d m e t a l - li n e a b s o r p t i o n f e a t u r e s . MNRAS000 , 1–19 (2017) Chen et al. k p c J J J J J J J J Normalized Flux d = H I H I M g II M g II S i II S i II S i III C III O V I O V I R e l a t i v e V e l o c i t y ( k m / s ) F i g u r e b . S a m e a s F i g u r e , bu t f o r t h e r e m a i n i n g e i g h t L R G h a l o s a t d = − k p c . F o r t h e L R G s a t d = nd k p c , n oa b s o r p t i o n f e a t u r e s a r e d e t ec t e d a nd - σ upp e r li m i t s a r e o b t a i n e d w i t h i n ± s p ec t r a l r e s o l u t i o n e l e m e n t , t h e w i nd o w m a r k e db y t h e t h i c k h o r i z o n t a l b a r . T h e L R G a t d = k p c i s t h e o n l y o n e o f a ll L y α - a b s o r b i n g L R G s i n o u r s a m p l e w i t hn o d e t ec t a b l e m e t a l - li n e a b s o r p t i o n f e a t u r e s . MNRAS , 1–19 (2017) G M i n M a ss i ve Q u i e s ce n t H a l o s Table 4.
Summary of LRG Halo Gas Properties
LRG Absorption Properties d v gas − gal dv c b H I W r (1215) W r (977) W r (1031) W r (1206) W r (1260) W r (2796)ID z LRG (kpc) z abs (km s − ) comp. (km s − ) log N (H I ) (km s − ) (m˚A) (m˚A) (m˚A) (m˚A) (m˚A) (m˚A)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)SDSSJ 094632.40+512335.9 0.4076 41.7 0.40701 −
126 all ... 17 . ± .
01 ... 1639 ±
47 926 ±
37 800 ±
27 688 ± <
420 ...1 − . +4 . − . . +0 . − . . +3 . − . ... ... ... ... ... ...2 0 . +0 . − . . ± .
01 12 . +0 . − . ... ... ... ... ... ...3 +54 . +4 . − . . +0 . − . . +3 . − . ... ... ... ... ... ...4 +132 . +10 . − . . +0 . − . . +9 . − . ... ... ... ... ... ...SDSSJ 140625.97+250923.2 0.4004 47.3 0.40040 0 all ... 18 . +0 . − . ... 1535 ±
40 610 ±
140 ... 433 ±
30 192 ±
39 431 ± − . +3 . − . . ± .
22 28 . +2 . − . ... ... ... ... ... ...2 − . +5 . − . . +0 . − . . +2 . − . ... ... ... ... ... ...3 0 . +0 . − . . +0 . − . . +1 . − . ... ... ... ... ... ...4 +82 . +4 . − . . +0 . − . . +3 . − . ... ... ... ... ... ...SDSSJ 111132.33+554712.8 0.4629 77.1 0.46352 +127 all ... 17 . ± .
01 ... 876 ±
27 401 ± ±
10 362 ±
21 ... ...1 − . +1 . − . . ± .
03 20 . +1 . − . ... ... ... ... ... ...2 − . +2 . − . . +0 . − . . +1 . − . ... ... ... ... ... ...3 0 . +0 . − . . ± .
01 14 . +0 . − . ... ... ... ... ... ...SDSSJ 080357.74+433309.9 0.2535 78.5 0.25330 −
48 1 0 . +2 . − . . ± .
05 42 . +3 . − . ±
18 175 ±
23 ... < < < . ± . . ± .
02 36 . +0 . − . ± >
474 ... 565 ±
19 730 ±
37 1129 ± −
52 all ... 18 . +0 . − . ... 1353 ±
27 689 ±
15 206 ±
14 500 ±
28 312 ±
16 584 ± − . +9 . − . . +0 . − . . +6 . − . ... ... ... ... ... ...2 0 . +0 . − . . +0 . − . . +0 . − . ... ... ... ... ... ...3 +95 . +2 . − . . +0 . − . . +3 . − . ... ... ... ... ... ...SDSSJ 112755.83+115438.3 0.4237 98.7 0.42461 +192 all ... 15 . ± .
06 ... 680 ±
31 185 ±
10 54 ±
12 90 ±
33 ... 62 ± − . +4 . − . . +0 . − . . +3 . − . ... ... ... ... ... ...2 0 . +3 . − . . +0 . − . . +2 . − . ... ... ... ... ... ...SDSSJ 124307.36+353926.3 0.3896 102.2 0.3896 ... ... 0 . < . a < < < < <
34 ...SDSSJ 155047.70+400122.6 0.3125 106.7 0.31282 +73 all ... 16 . ± .
04 ... 1264 ±
24 445 ±
27 72 ±
23 360 ± <
105 196 ± − . +3 . − . . ± .
06 39 . +1 . − . ... ... ... ... ... ...2 0 . ± . . +0 . − . . ± . . +4 . − . . ± .
05 54 . +7 . − . ... ... ... ... ... ...SDSSJ 024651.20 − . ± .
01 ... 1573 ±
34 365 ±
26 45 ± ±
42 146 ±
32 84 ± − . +6 . − . . ± .
06 59 . +3 . − . ... ... ... ... ... ...2 − . +0 . − . . +0 . − . . ± . − . +0 . − . . ± .
05 37 . +1 . − . ... ... ... ... ... ...4 +0 . +0 . − . . +0 . − . . ± . . +2 . − . . +0 . − . . +0 . − . ... ... ... ... ... ...SDSSJ 135727.27+043603.3 0.3296 125.9 0.32869 −
205 1 0 . ± . . ± .
002 18 . ± . ± < < < < < −
164 all ... 16 . +0 . − . ... 1198 ±
38 681 ± c ... 214 ±
60 73 ±
30 145 ±
181 0 . ± . . +0 . − . . +1 . − . ... ... ... ... ... ...2 +91 . +2 . − . . +0 . − . . +2 . − . ... ... ... ... ... ...3 +218 . +2 . − . . +0 . − . . +2 . − . ... ... ... ... ... ...SDSSJ 141307.39+091956.7 0.3584 149.2 0.3584 ... ... 0 . < . a < < < < <
17 ...SDSSJ 155304.32+354853.9 0.4736 155.9 0.47540 +366 1 0 . +1 . − . . ± .
04 22 . +1 . − . ±
76 206 ±
12 75 ± <
119 ... 49 ± . < . a < < < < <
21 ...SDSSJ 124409.17+172111.9 0.5591 160.1 0.55851 −
113 1 0 . +1 . − . . ± .
03 30 . +1 . − . b ±
14 ... ... ... 130 ± a The purpose of specifying the b value here is for generating the Voigt profile displayed in Figures 2 & 3 for the non-detections. b Ly α is not covered by the available COS spectrum; W r (1215) is inferred from the best-fit log N (H I ) and b . c Werk et al. (2013) reported W r (977) = 282 ±
63 m˚A for this feature. The discrepancy is likely due to a combination of low
S/N and uncertainties in the continuum normalization. M N R A S , ( ) Chen et al.
Our survey shows not only that H I absorbers are fre-quently seen at d <
160 kpc in LRG halos, but that a largefraction (7/16) of these absorbers are also optically thick toionizing radiation field with log N (H I ) > .
2, correspond-ing to an optical depth at the Lyman edge of τ >
1. Sucha high rate of incidence exceeds the expectation from ran-dom background counts. We estimate the expected numberof random Lyman limit systems (LLS) in the redshift rangesurveyed in the COS-LRG sample by adopting the num-ber density of z < ±
500 km s − around 16 LRGs at z ≈ .
4, we only expect to find between0.02 and 0.04 LLS over the entire COS-LRG sample.In addition, the observed large clustering amplitude ofLRGs (e.g., Zheng et al. 2007; Padmanabhan et al. 2007;Gauthier et al. 2009) also indicates that there are on av-erage more lower-mass galaxies clustered around the LRGsthan typical L ∗ galaxies on scales of Mpc and, through red-shift space distortions, absorbing gas in these correlated ha-los could contribute to the detection statistics around LRGs.No clustering measurements are available for LLS at low red-shifts. We therefore infer the contribution from correlatedgalaxy halos from observations of Mg II in the outskirts ofgalaxy clusters (Lopez et al. 2008; see also Muzahid et al.2017). Lopez et al. (2008) reported a factor of 4 − II absorbers ( W r (2796) > d < ±
500 km s − velocity window along 16 LRG sight-lines. This relatively small number of expected LLS fromrandom and correlated counts, in contrast to seven LLS de-tected at d <
160 kpc and | v gas − gal | < ∼
350 km s − from 16LRGs, provides a strong support for a physical connectionbetween the observed optically-thick gas and the LRGs.We present the spatial distribution of N (H I ) as a func-tion of d for the COS-LRG sample in Figure 4a. For com-parison, we also include the COS-Halos red galaxies (Thomet al. 2012) with updated N (H I ) from Prochaska et al.(2017). In Figure 4b, we show the distribution of N (H I )versus M star for both samples. No correlation is found be-tween N (H I ) and M star . Recall that COS-LRG sample isuniformly defined to have log M star / M (cid:12) > ∼
11 with a me-dian and dispersion of (cid:104) log M star / M (cid:12) (cid:105) COS − LRG = 11 . σ (cid:104) log M star / M (cid:12) (cid:105) COS − LRG = 0 . z = 0 . − . (cid:104) log M star / M (cid:12) (cid:105) COS − HalosRed = 10 . ± . σ (cid:104) log M star / M (cid:12) (cid:105) COS − HalosRed = 0 . z = 0 . − . N (H I ), a large fraction of the COS-LRG and the COS-Halosred samples contain optically-thick gas at d < ∼
100 kpc and ageneral declining trend of N (H I ) is also seen toward largerdistances.We determine the covering fraction and associated un-certainties of optically-thick gas in massive quiescent ha-los following the maximum-likelihood approach described inChen et al. (2010). The likelihood of observing an ensem-ble of LRGs with n showing associated LLS and m non- COS-Halos RedCOS-LRG ( a )( b ) Figure 4.
Top : The observed spatial distribution of N (H I ) as afunction of projected distance d for the COS-LRG sample (filledcircles), in comparison to those found for the COS-Halos redgalaxies (red diamonds; Thom et al. 2012). Downward arrowsrepresents a 2- σ upper limit to the underlying N (H I ) for haloswith no Ly α absorption line detected. The COS-Halos measure-ments have been updated based on the report of Prochaska etal. (2017). For two of the COS-Halos red galaxies, the avail-able Lyman series lines are all saturated and therefore onlylower limits to N (H I ) are available. These are shown as upwardarrows. Bottom : Distribution of N (H I ) versus M star for bothCOS-LRG and COS-Halos red samples. Recall that COS LRGsform a mass-limited sample of log M star > ∼
11 with a median of (cid:104) log M star / M (cid:12) (cid:105) COS − LRG = 11 . z = 0 . − .
55, while theCOS-Halos red sample includes predominantly lower-mass galax-ies with a median of (cid:104) log M star / M (cid:12) (cid:105) COS − HalosRed = 10 . z = 0 . − .
27. No correlation is found between N (H I ) and M star . detections is given by L ( κ LLS ) = (cid:104) κ (cid:105) n LLS [1 − (cid:104) κ (cid:105) LLS ] m . (1)The full COS-LRG sample probes the LRG halos out to d ≈
160 kpc (or d ≈ / R vir ). Based on the detectionsof LLS in seven of the 16 LRG halos, we estimate a meancovering fraction of optically-thick gas of (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
160 kpc. The error bars represent the 68% confidenceinterval. In the inner halos, where we detect LLS in five ofthe seven LRG halos, the mean covering fraction is (cid:104) κ (cid:105) LLS =0 . +0 . − . at d < ∼
100 kpc. We note that excluding five COS-Halos red galaxies does not change the results. We estimate
MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
160 kpc based on five detectionsof LLS around 11 randomly selected LRGs.The observed covering fraction of optically-thick gas inmassive quiescent halos at z ≈ . (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼ L ∗ and sub- L ∗ star-forming galaxies at z ≈ . (cid:104) κ (cid:105) LLS = 0 . +0 . − . observed at d < ∼
122 kpc in quasar host halos (e.g., Hennawi et al. 2006;Prochaska et al. 2013) and (cid:104) κ (cid:105) LLS = 0 . ± .
14 observedat d < ∼ R vir in halos hosting starburst galaxies at z ≈ . M h ∼ . M (cid:12) (e.g.,White et al. 2012), while the targeted starburst galaxies at z ≈ . M h ∼ M (cid:12) (e.g., Adelberger et al. 2005). Both populations reside inlower-mass halos at z > z ≈ .
4. Toreproduce the observed high gas covering fraction in thesedistant halos in cosmological simulations, it is necessary toinvoke energetic feedback due to either active galactic nuclei(AGN) in quasar host galaxies (e.g., Rahmati et al. 2015)or massive stars (e.g., Faucher-Gigu`ere et al. (2016). Thequiescent state of the LRGs makes either AGN or stellarfeedback a challenging scenario for explaining the frequentappearance of optically-thick gas in these massive halos.At the same time, we note the stark contrast betweenLRG halos at z ≈ . <
5% based on deep 21 cm surveys (Howk et al. 2017),while the covering fraction of Si
III absorbing gas is found toapproach unity from absorption spectroscopy of backgroundQSOs (Lehner et al. 2015). The lack of optically-thick gascoupled with a high covering fraction of heavy ions suggeststhat either the halo gas is highly ionized or the absorbingclouds are significantly smaller than the beam size ( ≈ ≈ . (cid:12) yr − (e.g., Barmby et al. 2006; Rahmaniet al. 2016) resides in a dark matter halo of M h ≈ M (cid:12) (e.g., Watkins et al. 2010; Sofue 2015). A lack of optically-thick gas would be qualitatively consistent with the observedlow SFR in M31 disk, but different from what is observedfor massive quiescent halos at z ≈ . n LLS (LRG) = cH (1 + z ) (cid:112) Ω M (1 + z ) + Ω Λ × (cid:90) ∞ L min dL φ ( L ) (cid:104) κ (cid:105) LLS ( πR ) , (2)where c is the speed of light, φ ( L ) is the LRG luminosityfunction, R gas is the radius of gaseous halos, and L min is theminimum luminosity of LRGs which is approximately 3 L ∗ for the COS-LRG sample, with a corresponding absolute i -band magnitude of M i = − .
6. Here we have assumed thatthe gas cross section does not vary significantly with lumi-nosity. Adopting the LRG luminosity function of Montero- Dorta et al. (2016), which is characterized by M i ∗ = − . φ ∗ = 7 . × − Mpc − , we find n LLS (LRG) ≈ . R gas = 160 kpc and (cid:104) κ (cid:105) LLS = 0 .
44. The observed numberdensity of LLS with log N (H I ) (cid:62) . n LLS = 0 . − . z < z < A primary motivation of the COS-LRG program is tounderstand the origin and significance of Mg II absorbersfound in LRG halos (e.g., Gauthier et al. 2009, 2010, Gau-thier & Chen 2011; Bowen & Chelouche 2011; Huang et al.2016). The COS-LRG sample is established without priorknowledge of the presence/absence of any absorption fea-tures in the halos, including metal absorption lines. It there-fore provides a necessary baseline calibration for characteriz-ing the significance of metal-line absorbers in these massivequiescent halos. The expanded spectral coverage from com-bining COS FUV and ground-based echelle spectra also pro-vides empirical constraints for multiple ionization species,from singly-ionized silicon and magnesium, to twice-ionizedcarbon and silicon, and to highly-ionized O .Figure 5 shows the rest-frame absorption equivalentwidth, W r versus d for Ly α and associated metal absorptionfeatures, including Mg II λ II λ III λ III λ VI λ I absorbers detected, associated metal absorption fea-tures are also detected with intermediate ions revealed byC III λ
977 and Si
III λ II and highly-ionized gas probed by O VI . For the three LRGswithout an H I absorber detected, no metal absorption linesare found to sensitive upper limits.SDSSJ 135726.27+043541.4 at z = 0 . d = 126kpc is the only LRG displaying no trace of metal absorptionlines to 2- σ limits of < .
09 ˚A, while showing a strong Ly-man absorption series with log N (H I ) = 17 .
5. The presenceof optically-thick gas combined with a complete absence ofionic absorption features places a stringent upper limit of gasmetallicity at < ∼ .
3% solar (see Zahedy et al. 2018 in prepara-tion), representing the lowest-metallicity gas detected near aluminous galaxy at intermediate redshift (cf. Ribaudo et al.2011b; Prochaska et al. 2017). This low metallicity is amongthe most metal-poor LLS at z < z = 0 .
33 also shows that intergalactic medium at z < I absorbers indicates a wide spread chemicalenrichment in these massive quiescent halos. An early resultof the COS-Halos survey is the O VI “bimodality” in galaxyhalos (Tumlinson et al. 2011): galaxies with higher specific-SFR exhibit on average stronger O VI absorption featuresin their halos than those with low specific-SFR. This phe- MNRAS , 1–19 (2017) Chen et al.
COS-LRG Ly α SiIIMgIILy α CIIISiIIILy α OVI ( a )( b )( c ) Figure 5.
Measurements of rest-frame absorption equiva-lent width, W r versus d for different absorption transitions.Panel ( a ) presents low-ionization Mg II λ II λ b ) presentsintermediate-ionization Si III λ III λ c ) presents high-ionization O VI λ α (black circles) are included in all three panels for compar-isons. For SDSSJ 124409.17+172111.9 at z = 0 . d = 160kpc, W r (1215) is shown as an open circle because Ly α is notcovered by the available COS spectrum and W r (1215) is inferredfrom the best-fit log N (H I ) and b . For all but one LRG haloswith Ly α detected, associated metal absorption lines are also de-tected with C III λ
977 and Si
III λ z = 0 . d = 126 kpc for which optically-thick gas ispresent with log N (H I ) = 17 . σ limits of < .
05 ˚A. nomenon is commonly attributed to a causal connection be-tween star formation in galaxies and stellar-feedback drivenchemical enrichment in galaxy halos (e.g., Ford et al. 2013;Suresh et al. 2015), in which case the observed heavy ele-ments in galaxy halos are ejected from galaxies by newlyformed young stars and the origin of heavy elements in qui-escent halos remains unknown. Alternatively, because lowspecific-SFR galaxies in the COS-Halos sample are also moremassive than their star-forming counterparts, the deficit ofstrong O VI absorption observed in these passive galaxies canalso be understood as due to oxygen being ionized to higherionization states (e.g., Oppenheimer et al. 2016). In thiscase, the presence/absence of O VI is not directly connectedto star formation or AGN feedback, but it requires the CGMto be pre-enriched to > ∼ /
10 solar at < R vir . Adoptingan equivalent width threshold of 0.1 ˚A, we estimate a mean
LINER-LRGCOS-LRGpassive-LRG
Figure 6.
Comparison of rest-frame absorption equivalent width, W r (2796) versus d for 11 COS LRGs (orange circles) withground-based echelle absorption spectra available and the SDSSLRG samples from Huang et al. (2016). Recall that roughly 10%of LRGs exhibit LINER features (e.g., Johnston et al. 2008; Hodgeet al. 2008). These LINER-like galaxies are shown in green cir-cles, while the remaining 90% of passive LRGs are shown in red.Available echelle spectra for the COS LRGs have allowed us touncover very weak Mg II absorption features of total integrated W r (2796) > ∼ .
05 ˚A that have been previously missed in SDSSspectra. While only three of the 11 COS LRGs have associatedMg II absorbers of W r (2796) > . II absorption detected to a 2- σ upper limitof W r (2796) ≈
50 m˚A. gas covering fraction of (cid:104) κ (O VI ) (cid:105) . = 0 . +0 . − . , consistentwith the finding for COS-Halos red galaxies (e.g., Werk etal. 2013). While quiescent galaxies exhibit on average fewerstrong O VI absorbers of column density log N (O VI ) > d <
160 kpc, the mean gascovering fraction of moderately strong O VI absorbers oflog N (O VI ) > . ≈
50% inquiescent halos (Johnson et al. 2015b).Of the 16 galaxies in the COS-LRG sample, five do nothave constraints for O VI absorption because the spectralregions are contaminated by either geocoronal emission orother strong features. For the 11 LRGs with available con-straints on the O VI absorption strength, six have associ-ated O VI absorbers of W r (1031) > .
05 ˚A, correspondingroughly to log N (O VI ) > . (cid:104) κ (O VI ) (cid:105) . = 0 . +0 . − . , extending previous findings tohigher-mass quiescent galaxies.Similarly, 11 of the 16 QSO-LRG pairs have opticalechelle spectra of the QSOs available for constraining the ab-sorption strengths of the associated Mg II lines. These highsignal-to-noise ( S/N ) and high spectral resolution echellespectra have allowed us both to uncover very weak Mg II absorption features of total integrated W r (2796) > ∼ .
05 ˚Athat have been previously missed in SDSS spectra and toresolve individual components (see Figure 3). While onlythree of the 11 COS LRGs have associated Mg II absorbers MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos of W r (2796) > . II absorption detected to a 2- σ upper limitof W r (2796) ≈
50 m˚A. Figure 6 presents a comparison of W r (2796) versus d between the COS-LRG sample and theSDSS LRG samples, both passive (red symbols) and LINER-like (green), from Huang et al. (2016).The mean covering fraction of Mg II absorbing gas at d < ∼
160 kpc is found to be (cid:104) κ (Mg II ) (cid:105) . = 0 . +0 . − . for a min-imum absorption equivalent width W = 0 . W = 0 . (cid:104) κ (Mg II ) (cid:105) . = 0 . +0 . − . at d <
160 kpc, comparable to what is observed in lower-mass, star-forming galaxies (e.g., Chen et al. 2010; Werk etal. 2013). As a cautionary note, none of the three “transpar-ent” halos with no detectable Ly α have optical echelle spec-tra available. Consequently, these LRGs do not contributeto the estimates of Mg II gas covering fraction here. Theestimated mean value above therefore likely represents anupper limit to the true value. Including the three “transpar-ent” halos as non-detections in Mg II places a lower limit at (cid:104) κ (Mg II ) (cid:105) . > .
43. Combining these limits, we thereforeinfer a mean covering fraction of Mg II absorbing gas in therange of (cid:104) κ (Mg II ) (cid:105) . ≈ . − .
55 in LRG halos.A surprising finding of the COS-LRG survey is the sig-nificantly higher incidence of C
III than Mg II absorption fea-tures around LRGs, indicating that chemically-enriched coolgas is more abundant in these massive quiescent halos thanpreviously thought from searches of associated Mg II absorp-tion. Twelve of the 16 LRGs in the sample exhibit associatedC III with W r (977) > . (cid:104) κ (C III ) (cid:105) . = 0 . +0 . − . for a minimum rest-frameabsorption equivalent width of W = 0 . d < ∼
160 kpc.The observed high covering fraction of C
III absorbing gasin LRG halos is comparable with what is reported for COS-Halos red galaxies with (cid:104) κ (C III ) COS − Halosred (cid:105) . = 0 . +0 . − . by Werk et al. (2013), which is also statistically consis-tent with what is seen in COS-Halos blue galaxies with (cid:104) κ (C III ) COS − Halosblue (cid:105) . = 0 . ± .
07 in the COS-Halossurvey. While many of the detected C
III lines are saturatedand therefore only lower limits to the underlying ionic col-umn density N (C III ) can be obtained, C
III , unlike O VI , doesnot appear to show a strong bimodality in galaxy halos.Similarly, Si III λ W r (1206)as for C III λ W = 0 . III , we conservativelyestimate a mean gas covering fraction of (cid:104) κ (Si III ) (cid:105) . =0 . − .
7, similar to what was found for COS-Halos red galax-ies (e.g., Werk et al. 2013).
The dominant presence of intermediate ions revealedby C
III and Si
III absorption constrains the temperature ofthe absorbing gas to be T < ∼ × K under either colli-sional or photo-ionization scenarios (see e.g., Oppenheimer& Schaye 2013). This is in contrast to T > ∼ . K expectedfor the hot halo as a result of gravitational heating, andsupports a multi-phase halo model, in which cool absorb-ing clouds are pressure-confined in a hot medium. Under a pressure equilibrium, this would imply a density contrastbetween cool clouds and the hot halo of ∼
100 : 1. Whiledense, cool clouds are expected to fall in a low-density hothalo, whether or not they can reach the galaxy to fuel starformation depends on the infall time relative to the disrup-tion time (e.g., Maller & Bullock 2004). Under a simple hy-drostatic equilibrium assumption, Gauthier & Chen (2011)calculated the minimum mass required for cool clouds to sur-vive thermal conduction from the surrounding hot medium.They found a minimum mass of M cl > ∼ M (cid:12) . On the otherhand, Huang et al. (2016) reported that the velocity dis-persion of Mg II absorbing gas relative to the host LRGsis suppressed in comparison to the expectations from virialmotion. These authors attributed the suppressed velocitydispersion to ram-pressure drag force and inferred a max-imum cloud mass of M cl ≈ × M (cid:12) . While combiningthese calculations suggests that the majority of cool cloudsformed at large distances may not be sufficiently massiveto survive the hot halo during infall, we note that the re-sults of these calculations depend sensitively on the modelassumptions of the properties of the hot halo and how coolclouds form (e.g., Brighenti & Mathews 2003). Recall thathigh column density gas is detected at d <
20 kpc aroundmassive lensing galaxies at z ≈ . M star > ∼ M (cid:12) in the local universe are observed tocontain primarily evolved stellar populations with little on-going star formation (e.g., Peng et al. 2010; Tinker et al.2013). The results from our study can therefore be appliedbroadly to massive halos of M star > ∼ M (cid:12) , but in the stillhigher-mass galaxy cluster regime observations have yieldedvery different results (see e.g., Yoon et al. 2012; Burchett etal. 2018). In addition, the LRGs also provide a unique sam-ple for identifying additional physical mechanisms, beyondstarburst and AGN feedback, that produce the observedoptically-thick clouds and heavy ions in galactic halos. Ajoint study of the ionization state of the gas and the galaxyenvironment is the next important step for obtaining keyinsights into these underlying feedback mechanisms. We have carried out an absorption-line survey of coolgas in halos around a mass-limited sample of LRGs withlog M star (cid:62)
11 at z = 0 . − .
55. The LRGs are selectedwith no prior knowledge of the presence/absence of any ab-sorption features. Our program is designed such that accu-rate and precise measurements of N (H I ) are possible basedon observations of the full hydrogen Lyman series. Our anal-ysis shows that not only H I absorbers are frequently seenat d <
160 kpc (or d < / R vir ) in LRG halos, but a largefraction (7/16) of these absorbers are also optically thick toionizing radiation field with log N (H I ) > .
2. In addition,all but one LRGs with detected H I absorption also exhibitassociated metal-line absorption features, indicating a wide MNRAS , 1–19 (2017) Chen et al. spread chemical enrichment in these massive quiescent ha-los. The main findings of our survey are summarized below.(1) The mean covering fraction of optically-thick gas is (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
160 kpc and (cid:104) κ (cid:105) LLS = 0 . +0 . − . at d < ∼
100 kpc in LRG halos. These numbers are consistentwith the high covering fraction of (cid:104) κ (cid:105) LLS = 0 . +0 . − . observedat d < ∼
80 kpc from L ∗ and sub- L ∗ star-forming galaxies at z ≈ . III and Si
III are themost prominent UV ionic features in these massive quies-cent halos with a mean covering fraction of (cid:104) κ (C III ) (cid:105) . =0 . +0 . − . for W r (977) (cid:62) . d <
160 kpc, compara-ble to what is seen for C
III in L ∗ and sub- L ∗ star-formingand red galaxies but exceeding what is typically observed forstrong Mg II or O VI absorbing gas around massive quiescentgalaxies.(3) While quiescent galaxies exhibit on average fewerstrong O VI absorbers of column density log N (O VI ) > d <
160 kpc, the mean gascovering fraction of moderately strong O VI absorbers oflog N (O VI ) > . (cid:104) κ (O VI ) (cid:105) . = 0 . +0 . − . .(4) The mean covering fraction of Mg II absorbing gasat d < ∼
160 kpc is found to be (cid:104) κ (Mg II ) (cid:105) . = 0 . +0 . − . for W r (2796) (cid:62) . W r (2796) (cid:62) . d <
160 kpc is (cid:104) κ (Mg II ) (cid:105) . = 0 . +0 . − . , comparable towhat is observed in lower-mass, star-forming galaxies.In conclusion, the COS-LRG survey has uncovered ahigh incidence of chemically-enriched cool ( T ∼ − K)gas in massive quiescent halos hosting LRGs at z ≈ .
4. Asignificant fraction of this chemically-enriched gas is opti-cally thick to ionizing photons. It shows that massive quies-cent halos contain widespread chemically-enriched cool gas.No clear presence of bimodality is found between blue star-forming and red quiescent galaxies in their H I gas contentor in the incidence of intermediate ions. ACKNOWLEDGMENTS
We thank Joop Schaye and Jonathan Stern for helpfuldiscussions. We thank Michael Rauch for his help in obtain-ing the MIKE spectra for some of the QSOs presented inthis paper. This work is based on data gathered under theHST-GO-14145.01A program using the NASA/ESA HubbleSpace Telescope operated by the Space Telescope ScienceInstitute and the Association of Universities for Research inAstronomy, Inc., under NASA contract NAS 5-26555. HWCand FSZ acknowledge partial support from NSF grant AST-1715692. FSZ acknowledges generous support from the Brin-son Foundation and the Observatories of Carnegie Institu-tion for Science during his visit at Carnegie Observatories asa Brinson Chicago–Carnegie predoctoral fellow. SDJ is sup-ported by a NASA Hubble Fellowship (HST-HF2-51375.001-A). This research has made use of the Keck ObservatoryArchive (KOA), which is operated by the W. M. Keck Obser-vatory and the NASA Exoplanet Science Institute (NExScI),under contract with the National Aeronautics and Space Ad-ministration.
REFERENCES
Adelberger, K. L., Steidel, C. C., Pettini, M., et al. 2005,ApJ, 619, 697Barmby, P., Ashby, M. L. N., Bianchi, L., et al. 2006, ApJ,650, L45Bergeron, J., & Stasi´nska, G. 1986, A&A, 169, 1Bernstein, R., Shectman, S. A., Gunnels, S. M., Mochnacki,S., & Athey, A. E. 2003, in Instrument Design and Per-formance for Optical/Infrared Ground-based Telescopes,eds. M. Iye, & A. F. M. Moorwood, Proc. SPIE, 4841,1694Blake, C., Collister, A., & Lahav, O. 2008, MNRAS, 385,1257Blanton, M. R. & Roweis, S. 2007, AJ, 133, 734Bowen, D. V., & Chelouche, D. 2011, ApJ, 727, 47Brighenti F., Mathews W. G., 2003, ApJ, 587, 580Burchett, J. N., Tripp, T. M., Wang, Q. D., Willmer,C. N. A., Bowen, D. V., & Jenkins, E. B. 2018, MNRAS,475, 2067Chabrier, G. 2003, PASP, 115, 763Charlton, J. C., Ding, J., Zonak, S. G., et al. 2003, ApJ,589, 111Chen, H.-W., Helsby, J. E., Gauthier, J.-R., et al. 2010,ApJ, 714, 1521Chen, H.-W., Gauthier, J.-R., Sharon, K., et al. 2014, MN-RAS, 438, 1435Chen, H.-W. 2017a, Outskirts of Distant Galaxies in Ab-sorption. In: Knapen J., Lee J., Gil de Paz A. (eds) Out-skirts of Galaxies. Astrophysics and Space Science Li-brary, vol 434. Springer, ChamChen, H.-W. 2017b, The Circumgalactic Medium in Mas-sive Halos. In: Fox A., Dav´e R. (eds) Gas Accretion ontoGalaxies. Astrophysics and Space Science Library, vol 430.Springer, ChamChurchill, C. W., & Vogt, S. S. 2001, AJ, 122, 679Conroy, C. 2013, ARA&A, 51, 393Eisenstein, D. J., Annis, J., Gunn, J. E., et al. 2001, AJ,122, 2267Faucher-Gigu`ere, C.-A., Feldmann, R., Quataert, E., et al.2016, MNRAS, 461, L32Faucher-Gigu`ere, C.-A. 2017, Observational Diagnostics ofGas Flows: Insights from Cosmological Simulations. In:Fox A., Dav´e R. (eds) Gas Accretion onto Galaxies. As-trophysics and Space Science Library, vol 430. Springer,ChamFord A. B., Oppenheimer B. D., Dav´e R., Katz, N.,Kollmeier, J. A., & Weinberg, D. H. 2013, MNRAS, 432,89Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman,J. 2013, PASP, 125, 306Gauthier J.-R., Chen H.-W., Tinker J. L., 2009, ApJ, 702,50—, 2010, ApJ, 716, 1263Gauthier J.-R., Chen H.-W., 2011, MNRAS, 418, 2730Hafen, Z., Faucher-Gigu`ere, C.-A., Angl´es-Alc´azar, D., etal. 2017, MNRAS, 469, 2292Hennawi, J. F., Prochaska, J. X., Burles, S., et al. 2006,ApJ, 651, 61Hodge, J. A., Becker, R. H., White, R. L., & de Vries, W. H.2008, AJ, 136, 1097Howk, J. C., Wotta, C. B., Berg, M. A., et al. 2017, ApJ,
MNRAS , 1–19 (2017)
GM in Massive Quiescent Halos MNRAS000