Deep galaxy spectroscopy has now provided perhaps the only unequivocal evidence for the inflow of gas toward galaxies beyond the local universe. While the earliest spectroscopic datasets probing cool gas kinematics “down the barrel” specifically targeted starbursting galaxies and typically reported only blueshifted (i.e., outflowing) absorption (Heckman et al. 2000, Rupke et al. 2005b, Martin 2005), later surveys achieving the requisite S/N have revealed numerous instances of redshifted metal-line absorption profiles toward a wide variety of host systems. Study of Na I D profiles in spectra of red-sequence galaxies at z ∼ 0.3 has suggested an incidence rate of inflow of ∼ 20%; i.e., similar to the rate of outflow detections for the same sample (Sato et al. 2009). Spectroscopy of Na I and the OH 119µm feature in the far-IR in small samples of nearby Seyfert and X-ray-bright AGN hosts points to a yet higher rate of redshifted cool gas absorption (∼ 40%; Krug et al. 2010, Stone et al. 2016). Such inflows may in fact be necessary to fuel the observed nuclear activity; however, the amount of mass carried in these flows remains poorly constrained, and current observations do not establish the present (or ultimate) location of the gas along the line of sight (Shi et al. 2016, for an exception to this generalization).
In the few years following these first detections, deep spectroscopy of Fe II and Mg II transitions in the rest-frame near-UV has finally revealed evidence for inflow onto “normal” star-forming galaxies at z > 0.3 (Rubin et al. 2012, Martin et al. 2012). The reported rate of incidence is low (< 10%), as blueshifted absorption tends to dominate the metal-line profiles in these spectra. Furthermore, Rubin et al. (2012) has presented evidence suggesting that inflow is more likely to be detected in these star-forming systems when they are viewed in an edge-on orientation. However, Martin et al. (2012) reported detections of inflow toward galaxies viewed over a wide range of orientations, and hence do not support this claimed dependence of inflow detection rate on viewing angle.
Such studies have been pivotal for furthering our understanding of the cycling of cool gas through galaxy environments. However, the evidence they offer is anecdotal rather than statistical. A complete, empirical picture of the baryon cycle must establish the incidence, mass, and morphology of gas inflow as a function of host galaxy stellar mass, star formation activity and history, and AGN luminosity. As all of the aforementioned samples are S/N-limited, they cannot assess such quantities regardless of their spectral resolution or the transitions they probe. Surveys of cool gas kinematics in samples selected to be complete to a given stellar mass limit (i.e., unbiased in their distributions of SFR and galaxy orientation) are required if we are to make substantive progress in our development of an empirical model of galactic gas flows.
At the same time, the observational challenges inherent in such an effort are significant. The works discussed in Section 3 surveying samples of ∼ 100−200 objects in the rest-frame near-UV together represent an investment of ∼ 17 nights on the Keck 1 Telescope (Rubin et al. 2012, Martin et al. 2012). A stellar mass-complete survey of a significantly larger sample would require a yet more extensive observing campaign. The ongoing SDSS-IV/MaNGA survey (Bundy et al. 2015) will facilitate a major advancement in our constraints on the incidence and morphology of inflow in the near term, as it will obtain high-S/N, spatially resolved spectroscopy of Na I in an unprecedented sample of ∼ 10,000 nearby galaxies. The MaNGA sample selection is carefully designed to be complete to a stellar mass logM* / M⊙ > 9, and its observing procedure ensures a minimum S/N per fiber of ∼ 6 per pix at 1.5 effective radii. Although its spectral coverage is limited to Ca II and Na I at relatively low velocity resolution (R ∼ 2000), its ∼ 1−2 kpc spatial resolution will enable detailed mapping of the incidence of cold, dusty gas inflow at speeds > 40 km s−1 in an extremely large galaxy sample.
To assess the mass and metallicity of these flows, deep, high resolution rest-frame far-UV spectroscopy will be required (Quider et al. 2009, 2010, Dessauges-Zavadsky et al. 2010). Such observations of samples of more than a few objects must await the next generation of wide-field multi-object spectrographs on 30m-class ground-based optical telescopes. As laid out in the Thirty Meter Telescope Detailed Science Case (Skidmore 2015), the prospective instrument WFOS will be capable of obtaining R ∼ 5000 spectroscopy of more than 100 (unlensed) LBGs at z ∼ 2−3 with RAB < 24.5 simultaneously. S/N > 30 will be achieved in just a few hours for these spectra, which will cover all of the ionic transitions discussed in the context of the Cosmic Eye (Section 4.1), including Lyα, O I, several Si II transitions, several Fe II transitions, Al III, Si IV, and C IV. Such a dataset will permit detailed constraints on the column densities, ionization states, metallicities, and mass of gas components arising in the ISM, outflows, and in accreting streams at high redshift.
Similarly detailed characterization of gas flows at z < 1.5 must await the next UV-sensitive space mission. A prospective high-resolution imaging spectrograph (France et al. 2016) conceived for the Large Ultraviolet/Optical/InfraRed (LUVOIR) surveyor NASA mission concept will not only access the important UV transitions discussed above, but will do so at a spatial resolution of ∼ 10−100 pc. This instrument will readily differentiate between material flowing inward in accreting streams from ongoing outflow and establish the mass and morphology of this accreting material. These capabilities will ultimately allow us to complete our empirical picture of the cycling of diffuse baryons through galaxy environments.