It is possible that metallicity can be used to determine the origins of absorbing gas observed around galaxies since outflows are expected to be metal-enriched while accreted gas should have lower metallicity (e.g., Shen et al., 2012). Gas accretion is expected to be metal poor but not purely pristine given that the first generations of Population III stars have likely enriched the gas to 10−4 Z⊙ by a redshift of 15−20 (e.g., Yoshida et al., 2004). In fact, there is a metallicity floor whereby it is rare to find absorption systems with metallicities much lower than 10−3 Z⊙ even out to z ∼ 5 (Prochaska et al., 2003; Penprase et al., 2010; Cooke et al., 2011; Battisti et al., 2012; Rafelski et al., 2012; Jorgenson et al., 2013; Cooke et al., 2015; Cooper et al., 2015; Fumagalli et al., 2016a; Lehner et al., 2016; Quiret et al., 2016). Although a few systems do have metallicities < 10−3 Z⊙ (Fumagalli et al., 2011b; Cooke et al., 2011; Cooke et al., 2015; Lehner et al., 2016) and possibly have Population III abundance patterns (Crighton et al., 2016). Cosmological simulations predict gas accretion metallicities should be between 10−3 − 10−0.5 Z⊙, which is dependent on redshift and halo mass (Fumagalli et al., 2011a; Oppenheimer et al., 2012; van de Voort & Schaye, 2012; Shen et al., 2013; Kacprzak et al., 2016), however this metallicity range does have some overlap with the metallicities of recycled outflowing gas.
There has been an abundance of studies that have identified galaxies with circumgalactic gas metallicity measurements in an effort to determine the source of the absorption. The general census shows that absorption systems near galaxies are either metal-poor with metallicities between −2 < [X/H] < −1 (Tripp et al., 2005; Cooksey et al., 2008; Kacprzak et al., 2010b; Ribaudo et al., 2011; Thom et al., 2011; Churchill et al., 2012; Bouché et al., 2013; Crighton et al., 2013; Stocke et al., 2013; Kacprzak et al., 2014; Crighton et al., 2015; Muzahid et al., 2015; Bouché et al., 2016; Fumagalli et al., 2016b; Rahmani et al., 2016) or metal-enriched with metallicities of [X/H] > − 0.7 (Chen et al., 2005; Lehner et al., 2009; Péroux et al., 2011; Bregman et al., 2013; Krogager et al., 2013; Meiring et al., 2013; Stocke et al., 2013; Crighton et al., 2015; Muzahid et al., 2015; Muzahid et al., 2016; Péroux et al., 2016; Rahmani et al., 2016). Determining the fraction of metal-rich and metal-poor systems is complicated since all the aforementioned studies have a range of observational biases due to the way the targets were selected – some selected from the presence of metal-lines. A clear complication of tracing the circumgalactic gas using metal-lines as tracers may bias you towards metal enriched systems. Ideally, the best way to avoid such a biases is to select absorption systems by hydrogen only.
Selecting only by hydrogen (but not necessarily with known galaxy hosts), has shown that the metallicity distribution of all Lyman limit systems below z < 1 appear to have a bi-modal distribution (Lehner et al., 2013; Wotta et al., 2016). The shape of Lyman limit systems (17.2 < log N(H i) < 17.7 in their study) metallicity bimodality distribution could be explained by outflows producing the high metallicity peak ([X/H] ∼ −0.3) while accreting/recycled gas could produce the low metallicity peak ([X/H] ∼ −1.9) (Wotta et al., 2016). Interestingly, between 2 < z < 3.5, the metallicity distribution of Lyman limit systems uni-modal peaking at [X/H] = −2, in contrast to the bimodal distribution seen at z < 1 (Fumagalli et al., 2011b; Lehner et al., 2016). Therefore it is likely that there exists a vast reservoir of metal-poor cool gas that can accrete onto galaxies at high redshift and outflows build up the circumgalactic medium at a later time. These results are discussed in detail in the chapter by Nicolas Lehner.
The overall knowledge of the metallicity distribution of the circumgalactic medium provides critical clues to the physics of gas cycles of galaxies. The bi-modal metallicity distribution is suggestive that we are observing both outflows and accretion, but our assumptions rely strongly on predictions from simulations, which still have some issues with modeling the circumgalactic medium since it is extremely feedback dependent. One way to determine if we are observing outflows and gas accretion is to combine our expectation of the gas geometry of accretion being along the minor axis and metal poor, while outflows are metal-enrich and ejected along the minor axis.
Preliminary work by Péroux et al. (2016), using nine galaxies, indicate that there is a very weak anti-correlation with metallicity and azimuthal angle. Figure 6 show the relative metallicity of the absorbing gas with respect to the host galaxy metallicity as a function of azimuthal angle, which are two independent indicators of gas flow origins. Note that in the figure a positive difference in metallicity indicates a circumgalactic medium metallicity lower than that of the host galaxy's HII regions (expected for metal-poor accreting gas) while a negative value indicates a higher circumgalactic medium metallicity than the host galaxy (expected for metal-enriched outflows). For the few objects shown in Figure 6, there is not clear correlation as expected under simple geometric and metallicity assumptions. Note that, different from expectations, there does not appear to be any high metallicity gas at high azimuthal angles. At low low azimuthal angles, there are a range of metallicity differences including negative values, which is unexpected for accreting gas. Given the few number of systems and only a weak anti-correlation, more systems are required to understand if there is a relation between the spatial location of the circumgalactic medium and metallicity. This is an active area of research and may be the most promising avenue to peruse in the future.
Figure 6. Metallicity difference between the host galaxy and absorber as a function of azimuthal angle. In this plot, accreted gas is expected to reside in the upper-left corner (for low/co-planer azimuthal angle and high metallicity difference), while outflowing gas should reside in the lower-right corner (high/minor axis azimuthal angle and metallicity similar to the host galaxy) as indicated by the red arrows. Note outflowing gas appears to be metal-poor, while accreting gas exhibits a range in metallicity. Additional observations are necessary to better relate metallicity and geometry in gas flows as only a minor anti-correlation is currently measured. Image from Péroux et al. (2016).