Luminous red galaxies (LRGs) uncovered in the SDSS exhibit little on-going star formation and display photometric and spectral properties resembling nearby elliptical galaxies (Eisenstein et al., 2001). They are characterized by a mean luminosity of ≈ 5 L* and a mean stellar mass of Mstar ≈ 3 × 1011 M⊙ at z ≈ 0.5 (e.g. Tojeiro et al., 2011), with corresponding dark matter halo mass of Mh ≈ 3 × 1013 M⊙ (e.g. Zheng et al., 2007, Blake et al., 2008, Padmanabhan et al., 2009), and therefore offer an ideal laboratory for studying the cool gas content in massive quiescent halos.
Roughly 10% of LRGs exhibit [O II] emission features (Eisenstein et al., 2003, Roseboom et al., 2006), suggesting the possible presence of on-going star formation in some of these massive galaxies. However, a more detailed examination of their spectral properties reveals unusually high [N II] / Hα together with low [O III] / [O II] ratios, both of which indicate the presence of low-ionization nuclear emission line regions (LINER) in these galaxies (e.g. Huang et al., 2016). Figure 1 shows stacked spectra of SDSS LRGs grouped into “passive” and [O II]-emitting subsamples. In both panels of Figure 1, the stacked spectra are characterized by prominent absorption features due to Ca II H and K, G-band, Mg I and Na I, along with LINER-like emission, namely strong [N II] / Hα ratios. The prominent absorption features indicate a predominantly old stellar population, whereas the LINER-like emission indicates the presence of additional ionizing sources in these galaxies. Although the origin of LINER-like emission is not fully understood, the observed slow decline in the spatial profile of LINER-like emission requires spatially distributed ionizing sources, rather than centrally concentrated active galactic nuclei (AGN), in these quiescent galaxies (e.g. Sarzi et al., 2006, Sarzi et al., 2010, Yan & Blanton, 2012, Singh et al., 2013, Belfiore et al., 2016). Likely sources include photo-ionization due to hot, post-asymptotic giant branch (post-AGB) stars or winds from Type Ia supernovae (e.g. Conroy et al., 2015). In the absence of on-going star formation, these LRGs also provide a unique sample for testing additional feedback mechanisms for quenching star formation in massive halos, including gravitational heating (e.g. Johansson et al., 2009) and winds from evolved stars and Type Ia supernovae (e.g. Conroy et al., 2015).
Figure 1. The quiescent nature of z ≈ 0.5 luminous red galaxies (LRGs) shown in stacked rest-frame optical spectra. Panel (a) shows the median stack of ∼ 1780 passive LRGs and panel (b) shows the median stack of ∼ 280 LRGs with [O II] emission detected (see Huang et al. (2016) for a detailed description of the sample definition and selection). Prominent absorption and emission lines are labeled in red. The corresponding 1-σ dispersion in each stack is shown in cyan at the bottom of each panel. In both stacked spectra, the spectral properties are characterized by prominent absorption features due to Ca II H and K, G-band, Mg I and Na I that indicate a predominantly old stellar population, as well as a relatively weak Balmer absorption series. The observed high [N II] / Hα emission ratios seen in both passive and [O II]-emitting LRGs, together with the low [O III] / [O II] ratios, suggest the presence of LINER in these LRGs, which could be attributed to the presence of underlying AGNs or evolved AGB stars in these galaxies.
A first step toward understanding the quiescent state of LRGs is to characterize the incidence and covering fraction of cool gas in their host halos. Over the past several years, absorption-line observations of background QSOs have revealed extended cool gas at projected distances d ≈ 10−300 kpc from LRGs at z ∼ 0.5. Specifically, the LRG–Mg II cross-correlation function displays an amplitude that is comparable to the LRG auto-correlation on small scales (≲ 450 comoving kpc), well within the virial radii of the halos. The large clustering amplitude on small scales suggests the presence of Mg II gas inside these massive quiescent halos (e.g. Bouché et al., 2006, Tinker & Chen, 2008, Gauthier et al., 2009, Lundgren et al., 2009). In addition, surveys of Mg II λλ 2796, 2803 absorption doublets in LRG halos have covered a non-negligible fraction of LRG halos with associated Mg II absorbers (e.g. Bowen & Chelouche, 2011, Gauthier & Chen, 2011, Gauthier et al., 2014, Zhu et al., 2014, Huang et al., 2016).
These Mg II absorption transitions are commonly seen at projected distances of d ≲ 100 kpc from star-forming galaxies (e.g. Lanzetta & Bowen, 1990, Steidel et al., 1994, Nestor et al., 2007, Kacprzak et al., 2008, Chen & Tinker, 2008, Chen et al., 2010a, Werk et al., 2013) and are understood to arise primarily in photo-ionized gas of temperature T ∼ 104 K (e.g. Bergeron & Stasińska, 1986, Hamann, 1997) and neutral hydrogen column density N(H I) ≈ 1018−1022 cm−2 (e.g. Churchill et al., 2000, Rao et al., 2006). The large rest wavelengths make the Mg II doublet a convenient probe of chemically enriched gas at z ≈ 0.3−2.3 in optical QSO spectra. In addition, the rest-frame absorption equivalent width Wr(2796) is found to increase proportionally with the number of individual components and the velocity spread of the components (e.g. Petitjean & Bergeron, 1990, Churchill et al., 2000, Churchill et al., 2003). Measurements of Wr(2796) therefore trace the underlying gas kinematics corresponding to ∼ 100 km s−1 per Angstrom. Several authors have proposed that strong Mg II absorbers of Wr(2796) > 1 Å originate in starburst driven outflows (e.g. Zibetti et al., 2007, Ménard et al., 2011) and that the large Wr(2796) is driven by non-gravitational motion in the outflowing media (e.g. Bouché et al., 2006). The presence of strong Mg II absorbers near quiescent galaxies is therefore particularly surprising.
Figure 2 compares the observed mean covering fraction of Mg II absorbing gas, ⟨κ⟩Mg II , in halos around typical L*- and sub-L*-type, star-forming galaxies (blue triangles) with what is seen in LRG halos (green and red symbols) at z ≈ 0.3−0.5. For a representative comparison across a broad luminosity (or mass) range, the gas covering fraction is determined within a fiducial halo gas radius, Rgas, that scales with galaxy luminosity according to Rgas ≈ 107 (LB / LB*)0.35 kpc (Kacprzak et al., 2008, Chen et al., 2010a). In addition, κ is estimated for halo clouds with absorption strength exceeding W0 = 0.3 Å, a detection threshold that is allowed by the quality of the absorption spectra. Previous studies have shown that typical L* galaxies have Rgas ≈ 130 kpc and Rgas scales with luminosity according to Rgas ∝ L0.35 (see Chen & Tinker, 2008, Chen et al., 2010a). Following this scaling relation, LRGs with a mean luminosity of ≈ 3.6 L* are expected to have Rgas ≈ 206 kpc.
Figure 2. Mass dependence of mean covering fraction of chemically-enriched cool gas in galaxy halos (Chen et al., 2010a, Gauthier & Chen, 2011, Huang et al., 2016). The mean covering fraction, ⟨κ⟩, is calculated based on observations of Mg II absorbing gas within a fiducial halo gas radius, Rgas. Rest-frame r-band magnitude is adopted here as a proxy for the underlying total stellar mass (see Liang & Chen, 2014). Star-forming galaxies at z ≈ 0.3 are shown in blue triangles, while luminous red galaxies (LRGs) at z ≈ 0.5 are shown in squares/circles. The covering fraction is determined for gas with absorption strength exceeding W0 = 0.3 Å, a detection threshold that is allowed by the quality of the absorption spectra. The horizontal bars represent the full range of Mr for galaxies included in each bin, and the vertical error bars represent the 68% confidence interval. In comparison to star-forming galaxies (Chen et al., 2010a), LRGs, being old and massive, exhibit a much reduced covering fraction of chemically-enriched cool gas as probed by the Mg II absorption features (Gauthier & Chen, 2011, Huang et al., 2016). However, observations also show a definitive presence of cool gas in these LRG halos, demonstrating that they are not completely devoid of star formation fuels (Huang et al., 2016). Furthermore, roughly 10% of LRGs exhibit [O II] emission features and LINER-like spectra. These [O II]-emitting LRGs display a slightly elevated cool gas covering fraction from passive LRGs but the mean value remains significantly lower than what is seen in lower-mass, star-forming galaxies (Huang et al., 2016). Similar trends in the mean gas covering fraction have also been reported for H I Lyα and high-ionization species probed by the O VI λλ 1031,1037 doublet transitions. Massive quiescent galaxies also consistently display non-zero covering fractions of Lyα and [O VI] absorbing gas (Tumlinson et al., 2011, Thom et al., 2012, Tumlinson et al., 2013, Werk et al., 2013).
Interpreting the rest-frame r-band absolute magnitude as a proxy for the underlying stellar mass (e.g. Liang & Chen, 2014), Figure 2 displays a strong mass dependence in the observed incidence and covering fraction of Mg II absorbing gas. Specifically, the observed Mg II covering fraction in LRG halos is more than four times lower than what is found in halos around lower-mass, star-forming galaxies. However, in spite of this substantially reduced cool gas content in massive quiescent halos, the distinctly non-zero Mg II gas covering fraction around LRGs demonstrates the definitive presence of chemically-enriched cool gas around evolved galaxies. Finally, it is interesting to note that while [O II]-emitting LRGs display a slightly elevated overall cool gas content than passive LRGs, the mean gas covering fraction remains under 20% (Huang et al., 2016).
Figure 2 is based on two separate studies designed to characterize the CGM properties of L* / sub-L* galaxies using ∼ 260 galaxies at z ≈ 0.25 (Chen et al., 2010a) and those of massive quiescent galaxies using ∼ 38000 LRGs (Huang et al., 2016), respectively. The COS-Halos survey was designed for a systematic study of gaseous halos over a broad range of mass and star formation history, and the sample contains 28 blue sub-L* galaxies with a median mass of ⟨ Mstar(blue) ⟩ ≈ 2 × 1010 M⊙ and 16 massive red galaxies with a median mass of ⟨ Mstar(red) ⟩ ≈ 1011 M⊙ (Tumlinson et al., 2013, Werk et al., 2013). While the uncertainties are large due to a smaller sample size, the COS-Halos galaxies confirm the observed mass dependence of Mg II gas covering fraction in Figure 2 (Werk et al., 2013).
A declining gas covering fraction from low-mass, star-forming galaxies to massive quiescent galaxies is also observed in H I and highly ionized species probed by the O VI λλ 1031,1037 doublet transitions. Specifically, the covering fraction of moderately strong Ly α absorbers of Wr(1215) > 0.1 Å declines from ≳ 90% around blue, star-forming galaxies (Tumlinson et al., 2013) to ≈ 40−50% around red passive galaxies (Thom et al., 2012), and the covering fraction of O VI absorbing gas decreases from ≈ 90% to ≈ 30% (Tumlinson et al., 2011, Werk et al., 2013). In both cases, massive quiescent galaxies consistently display non-zero covering fractions of Lyα and O VI absorbing gas.
Deep 21 cm and CO imaging surveys have also revealed that roughly 40% of elliptical galaxies in the nearby universe contain cool neutral gas (e.g. Oosterloo et al., 2010, Young et al., 2014). While on-going star formation is observed at low levels, ≲ 1 M⊙ yr−1, in many early-type galaxies of Mstar ≈ 5 × 1010 M⊙, it is rarely seen in more massive systems of Mstar ≳ 1011 M⊙ (e.g. Salim & Rich, 2010). Combining H I/CO imaging surveys of local elliptical galaxies and absorption-line observations of distant early-type galaxies at z ≈ 0.2−0.5 demonstrates that cool gas is indeed present in some, although not all, massive quiescent halos.
A declining cool gas fraction with increasing halo mass is expected in theoretical models that attribute the observed Mg II absorbers to infalling gas from either thermally unstable hot halos or the intergalactic medium (e.g. Maller & Bullock, 2004, Kereš et al., 2009). In the presence of cool gas, it is expected that these halo clouds could provide the fuels necessary for sustaining star formation in the LRGs. The lack of on-going star formation in these massive galaxies over such an extended cosmological time period (≳ 2 Gyr; Gauthier & Chen, 2011) therefore suggests a prolonged duty cycle for the underlying heat sources.