Next Contents Previous


Passive and/or quenched galaxies possess little if any cold gas in their ISM, and blaming the CGM merely relocates the problem: how and why do these massive galaxies that once possessed a cold ISM lose and not regain it? Presumably their dark matter halos continue to add mass, but the accompanying gas does not enter the ISM and form stars like it once did. How galaxies achieve this transition is a deep and abiding problem in astrophysics, and the array of possible mechanisms for consuming, removing, and/or heating cold gas are beyond the scope of our review. We address the phenomenon of quenching by considering the CGM as a factor in, and indicator of, the quenching process.

LRG : Luminous Red Galaxies

8.1. The Fate of Cold Accretion and The Problem with Recycling

The accretion of gas into halos, its heating to around Tvir, and eventual cooling and entry to the ISM was long the prevailing picture of galaxy fueling. In an important twist on this basic picture, Kereš et al. (2005) argued that star-forming galaxies are fed by “cold accretion” never reaches Tvir but entered a galaxy's disk via streams while remaining below T ∼ 105 K. Above log M / M ∼ 10.3−10.5 (or Mhalo ∼ 1012 M), the dark matter halo has sufficient mass, and the CGM enough pressure, to support a virial shock and suppress the cold mode. The coincidence of this mass with the stellar mass that divides star-forming from passive galaxies (Figure 2) drew great attention to this scenario e.g. Dekel & Birnboim (2006), leading to predictions that the halos of passive galaxies should possess little cold gas (Stewart et al., 2011).

The observational picture belies the clean transition seen in simulations and the stark division of observed star formation rates. While COS-Halos did find a dramatic difference in highly ionized O vi around star-forming and passive galaxies, the latter do not show as strong a deficit of CGM H i. As shown in Figure 11, the equivalent widths and covering fractions of H i do not drop as stellar mass increases across the range log M ≃ 10−11 (Thom et al., 2012). This is directly contrary to the expectation from, e.g. Stewart et al. (2011) that the covering fraction of strong H i should drop to nearly zero as galaxies transition to the hot mode of accretion. The inner CGM (< 50 kpc), however, is not well covered by these observations (Figure 4); it is possible that high pressure hot gas close to the galaxy prevents this cold material from accreting, as some models predict (Schawinski et al., 2014).

Figure 11

Figure 11. Three views of the CGM and quenching. Top: a trend in Lyα equivalent width over three decades in stellar mass from COS-Halos (Tumlinson et al., 2013, purple) and COS-Dwarfs (Bordoloi et al., 2014a, orange). As shown by Thom et al. (2012), the presence of H i around red, passive galaxies indicates that their halos are not devoid of cold gas. Middle: Mg ii from COS-Halos and MAGIICAT (Nielsen et al., 2016, green). Bottom: the galaxy SFR bimodality from Figure 2.

The presence of cool gas in the halos of massive red galaxies is now well-established by Mg ii studies. Gauthier, Chen & Tinker (2010) and Bowen & Chelouche (2011) found covering fraction of fc = 10−20% out to 100-200 kpc for > 1Å absorbers around LRGs. Using a sample of ∼ 4000 foreground galaxies at z = 0.5 − 0.9 from the zCOSMOS survey, Bordoloi et al. (2011) found that the Mg ii equivalent width for blue galaxies is 8–10 times stronger at inner radii (< 50 kpc) than for red galaxies, but even red galaxies possess evidence for cold gas. Using a new SDSS-based catalog of Mg ii QSO absorbers and LRGs, Zhu et al. (2014) mapped the mean profile out to ≫ 1 Mpc scales, and argue that the mean profile at this mass scale is even stronger than found by Bordoloi et al., extending at a detectable level out to 1 Mpc for LRGs. Johnson, Chen & Mulchaey (2015b) have pointed out that strong Mg ii absorbers are usually consistent with being bound to their host halos, meaning that the cold gas is contained with the dynamical influence of the galaxy.

From a theoretical perspective, the quenching of galaxies is still a significant unsolved problem. Star formation must be curtailed, and later accretion and cooling of gas must be suppressed indefinitely to explain how galaxies remain passive for > 6 Gyr (Gallazzi et al., 2008). Theories vary in how they accomplish this: some models artificially truncate star formation based on halo mass (Somerville & Davé, 2015), while others suppress the star-forming fuel by heating the CGM itself (e.g., Gabor et al., 2010, Gabor & Davé, 2012). Thus the CGM itself can be the proximate cause of quenching, even if the source of CGM heating is not yet identified. Unfortunately models that manipulate the CGM directly cannot be tested against CGM observations, or at least, they must be modified somehow to recover the cold gas seen in passive galaxy halos.

By contrast, models that include self-consistent subgrid treatments of feedback, whether “thermal” (Schaye et al., 2015), “mechanical” (Choi et al., 2015), or a combination of thermal, mechanical, and radiative (Vogelsberger et al., 2014) can be compared to CGM observations as tests of their success. As an example, the mechanical feedback model implemented by Choi et al. (2015) performed better than the “standard” (Springel, Di Matteo & Hernquist, 2005, Di Matteo et al., 2008) thermal feedback model in both suppressing galaxy formation and reducing the surface density of gas in the CGM by factors of 3–10 at 10–100 kpc.

Suresh et al. (2015) addressed quenching using the Illustris simulations, which are tuned to the observed M / Mhalo and galaxy metallicities, but not the CGM. In Illustris, “thermal” AGN feedback is deposited locally, inside the galaxy, when the SMBH is in its energetic “quasar” mode. But in the ∼ 90% of the time when the SMBH is accreting quiescently, its “radio mode” feedback is deposited non-locally as thermal energy over 100 kpc scales. This amounts to direct heating of the CGM, shifting cold gas to intermediate temperatures showing more O vi, and otherwise warm gas to high temperatures showing O vii and O viii. The net effect is that the Tumlinson et al. (2011) trend of strong O vi around star forming galaxies and weak O vi around passive galaxies is recovered. The “cold” CGM is reduced, but not completely destroyed. To be consistent, any visible effects of feedback would need to persist even when the AGN is not active, as the COS-Halos galaxies in question are not AGN at the time we observe them. The EAGLE simulations presented by Oppenheimer et al. (2016b) show a similar conclusion with models of thermal feedback and non-equilibrium cooling: at higher mass, with more feedback, O vi is suppressed and the cold gas is depleted but not completely destroyed. These feedback effects force behaviors that generally resemble the data: they suppress star formation to create a red sequence, they force net gas loss from the inner CGM by heating gas that then bouyantly rises, and they shift the balance of gas ionization toward higher temperatures and higher ions.

Despite these advances, the basic paradox of quenching remains: what happens to the halos of passive galaxies to quench their star formation, keep it quenched, and yet leave cold gas present in their halos? If passive galaxies possess cold gas and are not using it, can we be sure of the (naively obvious) conclusion that star-forming galaxies are using the diffuse gas they possess? Moreover, if the bulk of star formation at low-z comes from recycled accretion, then to understand both how galaxies get their gas and how galaxies quench, we must understand how both the internal and external fuel supplies are shut off.

8.2. The CGM of AGN and Quasars

If feedback from AGN is effective at quenching their star formation and their cold CGM in simulations, it naturally suggests that this effect will be visible in the gaseous halos of galaxies with ongoing AGN activity. While hard radiation fields of AGN may leave distinctive ionization signatures in halo gas even long after the AGN fades (Keel et al., 2012, Oppenheimer & Schaye, 2013a), studies like COS-Halos with subsamples of passive galaxies have excluded active AGN for the most part, and even so have not seen any apparent signs of AGN effects on the CGM. No published study has systematically examined background QSO/foreground AGN pairs, though there is one such study underway with Hubble / COS 2.

At z > 2, the “Quasars Probing Quasars” (QPQ) program has seen clear evidence that galaxies hosting bright quasars show greatly enhanced gas budgets in H i and low ions (Prochaska, Lau & Hennawi, 2014) though less excess in the high ions. This enhancement of neutral and low-ionization gas hints at a larger accretion rate for these robustly star-forming galaxies. AGN may even yield a net gain of cold gas in the CGM Faucher-Giguere et al. (2016). The Lyα blobs observed at z > 2 may be gas accreting on to galaxies, with radiation powered by gravitational infall (Goerdt et al., 2010), though these data may be more consistent with illumination from buried AGN (Prescott, Martin & Dey, 2015). The higher gas masses only exacerbate the problem of feedback and quenching — there is more gas to be removed, and it is still not clear how that gas is removed or heated and accretion suppressed thereafter. Future work should focus on following such galaxies down through cosmic time as their QSOs fade, star formation is quenched, and the galaxies later evolve passively. Post-AGN and post-starburst galaxies should be examined for CGM gas as much as is practical. Understanding this process is critical to properly understanding the role of the CGM in creating or reflecting the birth of the red sequence.

[Data in Need of More Theory]

  1. Are there any clean observational tests or theoretical discriminants between the various heuristic models of feedback?
  2. Are there self-consistent models of quenching that produce a red sequence of galaxies and yet leave a significant mass of cold CGM? How is the remaining cold gas kept from accreting?
  3. What do the detailed kinematic profiles of the multiphase suite of absorbing ions tell us about the physical and dynamic structure of the CGM?

[Theory in Need of More Data]

  1. What is the mass and composition of the CGM at high-redshift and in low-z M < 1010 M galaxies, and how do these constrain galaxy evolution models?
  2. What is the small-scale density and kinematic structure of the CGM, and what does it tell us about the physics?
  3. What does the CGM do as galaxies quench? Does cool, neutral gas extend into the inner CGM of passive galaxies?
  4. Where are the metals that are still missing from the census? What are the elemental abundance ratios in CGM gas, and how do they depend on the galaxy's mass and star formation history?

2 Back.

Next Contents Previous