Figure 3 clearly indicates that despite being in a quiescent state, halo gas around massive LRGs has been enriched with heavy elements out to the virial radius. The chemical enrichment history in these massive halos, which can be broadly characterized by (i) gas metallicity and (ii) relative abundance pattern, offers independent clues for the origin of the observed cool halo gas. Specifically, gas metallicity quantifies the overall heavy element production level, and nearby galaxies are observed to follow a mass–metallicity relation with more massive galaxies displaying on average higher metallicities both in stars (Gallazzi et al., 2005, Kirby et al., 2013) and in the ISM (Tremonti et al., 2004, Lee et al., 2006). Therefore, a natural expectation is that gas ejected from massive galaxies should be highly metal-enriched, while gas ejected from low-mass satellites should exhibit a lower metallicity. Gas accreted from the IGM should contain still lower metallicities (e.g. Lehner et al., 2013, Kacprzak et al., 2014). However, uncertainties arise as a result of poorly understood processes involving mixing and transport of heavy elements (e.g. Tumlinson, 2006). If mixing is effective as metal-enriched materials propagate into the low-density, and presumably lower-metallicity, halo gas, then the overall metallicity is expected to be reduced. If mixing is ineffective, then large variations in chemical abundances are expected in single objects (e.g. Scale & Elmegreen, 2004). Consequently, metallicity provides only a weak indicator for possible origins of chemically-enriched cool gas in LRG halos.
On the other hand, the relative abundance pattern presents a fossil record for the primary source of heavy element production. Specifically, α-elements (such as O, Mg, Si, and S) are primarily produced in massive stars and core-collapse supernovae (SNe), while a significant fraction of heavier elements such as Fe and Ni are produced in Type Ia SNe over longer timescales (e.g. Tsujimoto et al., 1995, Nomoto et al., 2006). The observed [α/Fe] relative abundance ratio therefore constrains the relative contributions of core-collapse SNe and SNe Ia to the chemical enrichment history (Tinsley, 1979, Gilmore & Wyse, 1991, Tolstoy et al., 2003). Both distant galaxies (Wolfe et al., 2005) and nearby evolved stars (Matteucci, 2014) display an α-element enhanced abundance pattern, indicating a chemical enrichment process driven by core-collapse SNe in the early universe. At the same time, the ISM of nearby elliptical galaxies (e.g. Mathews & Brighenti, 2003, Humphrey & Buote, 2006) and intracluster medium (e.g. de Plaa et al., 2007) display super-solar Fe/α abundance ratios that reveal increased contributions from SNe Ia in the circumgalactic gas of massive, evolved galaxies. Relative abundance measurements between Fe and α-elements therefore provide a powerful constraint for the origin of chemically-enriched gas around galaxies.
At z ≳ 0.3, a series of Fe II absorption transitions are observable on the ground, which enable an accurate measurement of Fe II column density N(Fe II) on a component by component basis (e.g. Rigby et al., 2002, Narayanan et al., 2008, Zahedy et al., 2016a). Similarly, the Mg II doublet transitions enable accurate Mg II column density measurements N(Mg II) for all but a few saturated components. In principle, the total elemental abundances can be determined from the observed N(Fe II) and N(Mg II), if the ionization fractions of these ions are known. Gas metallicity can also be derived if the total hydrogen content can be inferred from measurements of the neutral hydrogen column density, N(H I). In practice, observations of N(H I) at low redshifts require UV spectroscopy carried out in space, and have only been done for a relatively small sample of galaxies (e.g. Chen & Prochaska, 2000, Werk et al., 2014). Furthermore, uncertainties in the derived ionization fractions of individual ions can be large and difficult to estimate under simple ionization models.
Figure 6. Enhanced Fe abundances in halos around massive quiescent galaxies in comparison to those around star-forming galaxies (Zahedy et al., 2016a, Zahedy et al., 2016b). Left: Expected ionization fractions of Mg+ and Fe+ ions from Cloudy photo-ionization calculations (Ferland et al., 2013) for photo-ionized gas of temperature T = 104 K and 0.1 solar metallicity in optically-thin (with neutral hydrogen column density N(H I) = 1015 cm−2) and optically-thick (with N(H I) = 1019 cm−2) regimes (Zahedy et al., 2016a). Changing metallicity does not alter the results. An updated version of the Haardt & Madau (2001) ionizing radiation field at z = 0.5 is adopted for computing the ionization parameter U at different gas density nh displayed at the top of the panel. These calculations demonstrate that, independent of gas metallicity, the observed N(Fe II) / N(Mg II) in the right panel represents a lower limit to the underlying (Fe/Mg) relative abundances. Right: Observed mean Fe II to Mg II column density ratio versus projected distance based on absorption spectroscopy of halo gas around quiescent (red circles) and star-forming (blue stars) galaxies at ⟨ z ⟩ ≈ 0.5. For comparison, ISM measurements based on x-ray observations of 19 nearby elliptical galaxies (Humphrey & Buote, 2006) are also included in the left panel, and the dotted line shows the solar value. The fractional contribution of Type Ia supernovae to the chemical enrichment history in the inner 60 kpc of massive quiescent halos is found to be similar to what is observed in the solar neighborhood.
Observations of the relative abundance ratio between Mg+ and Fe+ ions are particularly useful, because Mg+ and Fe+ share similar ionization potentials (15 eV and 16.2 eV, respectively) and are the dominant ionization states of the respective elements in both neutral and cool photo-ionized gas. While the elemental abundances of iron and magnesium may be uncertain, the relative abundance ratio, (Fe/Mg) can be inferred with high confidence from the observed N(Fe II) / N(Mg II following log (Fe / Mg) = log N(Fe II) / N(Mg II) − log (fFe+ / fMg+), where fFe+ is the fraction of Fe in the singly ionized state and fMg+ is the fraction of Mg in the singly ionized state. The ratio of ionization fractions, fFe+ / fMg+, estimated based on a suite of photo-ionization calculations is shown in the left panel of Figure 6. The photo-ionization models are computed using the Cloudy software (Ferland et al., 2013). These models assume a photo-ionized gas of temperature T = 104 K and 0.1 solar metallicity in optically-thin (with neutral hydrogen column density N(H I) = 1015 cm−2) and optically-thick (with N(H I) = 1019 cm−2) regimes (Zahedy et al., 2016a). An updated version of the Haardt & Madau (2001) ionizing radiation field at z = 0.5 is adopted for computing the ionization parameter U (defined as the number of ionizing photons per atom) at different gas density nh displayed at the top of the panel. The ionization fraction of Fe+ remains roughly equal to that of Mg+ in the optically-thick regime and lower in optically-thin gas for the full range of U explored. Consequently, log (fFe+ / fMg+) ≲ 0 and the observed N(Fe II) / N(Mg II) marks a lower limit to the underlying Fe to α-element abundance ratio:
Experimenting with different gas metallicity does not change the predicted ionization fraction, and accounting for differential dust depletion of iron and magnesium would further increase the inferred [Fe/Mg] (Savage & Sembach, 1996). This exercise demonstrates that useful empirical constraints for the relative Fe and Mg abundances can be obtained even in the absence of accurate measurements of N(H I) and gas metallicity.
A recent study utilizing multiply-lensed QSOs for probing gas at small projected distances, d ≲ 20 kpc or ≈ 1−2 effective radii re, from the lensing galaxies has uncovered important new clues for the origin of chemically-enriched cool gas in massive halos (Zahedy et al., 2016a). These lensing galaxies at z = 0.4−0.7 share similar properties concerning both the quiescent state and halo mass scales (Mh ≳ 1013 M⊙), but display distinct absorption-line profiles between different lensing galaxies and between different sightlines of individual lenses. The apparent large scatter in the observed absorption profiles is consistent with the large scatter displayed in Figure 3 and discussed in Section 3. Most interestingly, all Mg II absorbers detected near these lensing galaxies are strong and resolved into 8−15 individual components over a line-of-sight velocity range of Δv ≈ 300 − 600 km s−1. The Mg II absorption is accompanied with even stronger Fe II absorption with matching kinematic profiles. Comparing the relative absorption strengths between individual components also yields uniformly large N(Fe II) / N(Mg II) ratios over the full range of velocity spread, Δv, with a median of ⟨ log N(Fe II) / N(Mg II) ⟩ ≈ 0 and a scatter of < 0.1 dex.
Following Equation (3), the observed ⟨ log N(Fe II) / N(Mg II) ⟩ ≈ 0 naturally leads to a super solar Fe/Mg abundance ratio near these massive lensing galaxies. This is in stark contrast to an α-element enhanced chemical composition found in young, star-forming galaxies (Dessauges-Zavadsky et al., 2004, Crighton et al., 2013, Fox et al., 2014, e.g.) and see Zahedy et al. (2016a) for a detailed comparison, and clearly indicates different origins between gas associated with star-forming galaxies and with massive quiescent galaxies.
Previous studies have shown that the spatial distribution of SNe Ia in nearby early-type galaxies follows the stellar light out to r ∼ 4 re (e.g. Förster & Schawinski, 2008) and that the ISM of these massive quiescent galaxies exhibits an iron-enhanced abundance pattern (e.g. Mathews & Brighenti, 2003, Humphrey & Buote, 2006). Therefore, it seems likely that the observed Fe-rich gas at d ≈ re from the lenses originates in the ISM of these massive galaxies, where SNe Ia play a dominant role in driving the observed large velocity width and Fe-rich abundance pattern (Zahedy et al., 2016a). If the ISM of massive galaxies at intermediate redshift is locally enriched by SNe Ia, then the observed Fe/Mg is expected to decline with increasing projected distance.
To test this hypothesis, a sample of 13 massive quiescent galaxies (including two lensing galaxies) and 14 star-forming galaxies at intermediate redshifts (⟨ z ⟩ ≈ 0.5) has been assembled (Zahedy et al., 2016b). These galaxies were selected to have absorption spectra of background QSOs at d ≈ 10−400 kpc available for constraining the radial profile of Fe/Mg. The right panel of Figure 6 displays the mean ⟨N(Fe II) / N(Mg II)⟩ averaged over all individual components per halo versus d for both star-forming galaxies (blue stars) and the quiescent galaxy population (red circles). Although the sample is still small, the distinction between quiescent and star-forming halos is already apparent in Figure 6. While the mean Fe/Mg ratio is consistent with an α-element enhanced pattern in the outer halo at d > 100 kpc for both quiescent and star-forming galaxies, halo gas at d < 100 kpc from quiescent galaxies exhibits an elevated iron abundance in comparison to star-forming galaxies.
The observed iron enrichment level in the inner halo of z ≈ 0.5 quiescent galaxies is consistent with the solar value, similar to what is observed in the ISM of nearby elliptical galaxies (e.g. Mathews & Brighenti, 2003, Humphrey & Buote, 2006) and in the intracluster medium (e.g. de Plaa et al., 2007). The radial decline of the Fe/Mg relative abundances supports the hypothesis that the gas is locally enriched by SNe Ia. The minimum fractional contribution of SNe Ia to the chemical enrichment in the inner halos of massive quiescent galaxies is found to be fIa ≈ 15−20% based on the expected yields for Type Ia and core-collapse SNe (Iwamoto et al., 1999) and the observed Fe/Mg ratio at d < 100 kpc.