[C ii] 158-μm emission from the host galaxies of damped Lyman-alpha systems

See allHide authors and affiliations

Science  24 Mar 2017:
Vol. 355, Issue 6331, pp. 1285-1288
DOI: 10.1126/science.aal1737

Identifying the hosts of quasar absorbers

If the line of sight from Earth to a distant quasar passes through foreground material, some of the quasar's light is absorbed. If a galaxy-sized quantity of gas intervenes, it forms a damped Lyman α system (DLA), visible as absorption lines in the quasar spectrum. Using the Atacama Large Millimeter/Submillimeter Array, Neeleman et al. observed two quasars with known DLAs. They detected emission from gas and dust in two foreground galaxies associated with the DLAs and were able to measure their star-formation rates. Combining these different tracers of DLAs will help us understand how galaxies evolved in the early universe.

Science, this issue p. 1285


Gas surrounding high-redshift galaxies has been studied through observations of absorption line systems toward background quasars for decades. However, it has proven difficult to identify and characterize the galaxies associated with these absorbers due to the intrinsic faintness of the galaxies compared with the quasars at optical wavelengths. Using the Atacama Large Millimeter/Submillimeter Array, we report on detections of [C ii] 158-μm line and dust-continuum emission from two galaxies associated with two such absorbers at a redshift of z ~ 4. Our results indicate that the hosts of these high-metallicity absorbers have physical properties similar to massive star-forming galaxies and are embedded in enriched neutral hydrogen gas reservoirs that extend well beyond the star-forming interstellar medium of these galaxies.

Gas plays a crucial role in the formation of galaxies by providing the fuel for star formation. However, the initial overdensities of baryons and dark matter from which galaxies form do not contain enough gas to sustain the observed star-formation rate (SFR) of galaxies (1, 2). Galaxies must therefore accrete gas from their surroundings. A comprehensive understanding of the accreting gas—in particular, neutral hydrogen (H i) gas, which is a crucial component (2)—is thus critical for understanding the formation and evolution of galaxies. Unfortunately, observing H i gas in emission from galaxies at cosmological distances is challenging because of the intrinsic weakness of its most reliable tracer, the H i 21-cm hyperfine structure line. This line is difficult to detect at even moderate redshifts with today’s telescopes (3), and at high redshifts (z ≥ 1), we must rely on studying H i gas in absorption.

As light from a bright background source, such as a quasar, travels toward us, it encounters pockets of H i gas, which imprint a characteristic absorption signature in the spectrum due to absorption at the frequency of the redshifted Lyman-alpha (Ly-α) line. The strongest H i absorption features in quasar spectra have H i column densities ≥2 × 1020 cm–2 and are known as damped Ly-α absorbers (DLAs) because their Ly-α absorption line profiles show distinctive damping wings (4). Observations of DLAs have revealed a wealth of information on the properties of the absorbing gas, including its kinematical signatures (5), metal enrichment (6), dust content (7), and temperature (8). Moreover, DLAs have been intimately linked to galaxies through direct imaging studies at intermediate redshifts (9, 10), scaling relations (11, 12), and cosmological simulations (13, 14). Surveys of DLAs therefore provide a powerful means to study H i gas surrounding high-redshift galaxies.

Unfortunately, directly detecting the starlight from the foreground galaxy that is associated with the DLA is challenging at optical wavelengths, owing to the presence of the much brighter background quasar (15). As a result, despite many recent searches (10, 1618), there are only a handful of high-redshift DLAs whose host galaxies have been identified by imaging studies. As such, there is little direct observational data on the nature of the galaxies that give rise to DLAs at different redshifts, including their mass, SFR, and the extent of the neutral H i gas surrounding the galaxy. At the same time, extensive data are available from deep optical and near-infrared surveys on the stellar properties of luminosity-selected galaxies (1921), but little is known about the gas surrounding these galaxies. A crucial link is therefore missing in connecting high-redshift galaxies with the gaseous envelopes that shape their evolution.

We report here on a search for the galaxies associated with two DLAs at z ~ 4 using the Atacama Large Millimeter/Submillimeter Array (ALMA). At submillimeter wavelengths, the background quasars emit little radiation, enabling a search for the host galaxy of the DLA at close angular separation (impact parameter) from the quasar. The DLAs toward quasars SDSS J081740.52+135134.5 and SDSS J120110.31+211758.5 were selected from a large sample of DLAs because of their higher-than-average metal content (6). We used ALMA to carry out a search for emission in both the singly-ionized carbon fine-structure line at a rest wavelength of 157.74 μm ([C ii] 158 μm) and the far-infrared (FIR) dust continuum from the two DLA host galaxies (22). The [C ii] 158-μm line is expected to be the strongest FIR line from galaxies at these redshifts (23), in part because it is the primary coolant of cold H i gas (24) and because it is the strongest observed FIR line in the local universe (25).

In each of the ALMA observations, we detected a >6σ emission feature at the frequency of the redshifted [C ii] 158-μm line at z = 4.2601 and 3.7978, which is offset from the quasar position by 6.2′′ and 2.5′′, for SDSS J081740.52+135134.5 and SDSS J120110.31+211758.5, respectively (Fig. 1). The redshifts of the DLAs, as measured from low-ionization atomic absorption lines (e.g., singly ionized iron and silicon), and the [C ii] 158-μm emission (Fig. 2) are within 100 km s–1 of each other (Fig. 2). The close proximity in both redshift space and angular separation on the sky indicates that the [C ii] 158-μm emission must come from a galaxy associated with the DLA.

Fig. 1 400-GHz continuum and [C ii] 158-μm emission from two damped Ly-α absorber (DLA) fields.

(A and B) The ≈400-GHz continuum emission from the regions surrounding two quasars (black stars). Black contours begin at 3σ and increase by Embedded Image; dashed contours indicate negative values. Gray contours are drawn at increments of 25σ. The axes give the relative physical (proper) distance at the DLA redshifts (i.e., z = 4.2584 and 3.7975 for SDSS J081740.52+135134.5 and SDSS J120110.31+211758.5, respectively). (C and D) Mean flux density over the full [C ii] 158-μm line profile displayed in Fig. 2, A and B, for a smaller region centered on the identified DLA host galaxies. No other emission lines are detected in these fields. The line contours begin at 3σ, with each subsequent contour increasing by Embedded Image. The size of the synthesized beam is shown in the bottom left of each panel. The dashed line is the measured major axis of the galaxy (22). QSO, quasar (quasi-stellar object).

Fig. 2 Emission and absorption spectra from the host galaxies and DLAs.

(A and B) [C ii] 158-μm emission profile for the galaxy hosts identified with two high-redshift DLAs (Fig. 1). The 1σ uncertainties are shown in gray. The zero point of the velocity scale was chosen to correspond with the strongest absorption feature of the DLAs [(C) and (D)]. The absorption profiles are for two representative low-ionization elements, singly ionized silicon and singly ionized iron, which trace the bulk of the metals in the absorbers. The grayed-out region in (C) is a sulfur absorption line (S ii λ1259 Å). The agreement in redshift and width of the absorption and emission lines indicate that the [C ii] 158-μm emission is from the DLA host galaxy.

Our [C ii] 158-μm images are sensitive to the kinematics of the galaxy, as any gas motion will yield a Doppler shift in the observed [C ii] 158-μm line frequency. Because the emission is only barely resolved spatially, the kinematic signature of the gas is best described by a position-velocity (p-v) diagram (Fig. 3). For ALMA J081740.86+135138.2, the shape of the emission is indicative of rotation, which is corroborated by the “double-horned” profile seen in the [C ii] 158-μm emission spectrum (Fig. 2). From the magnitude of the rotation, we estimate a lower limit to the dynamical mass of 6 × 1010 solar masses Embedded Image for this galaxy (22). For ALMA J120110.26+211756.2, the p-v diagram suggests more complicated gas dynamics, corroborated by the spatial offset between the redshifted and blueshifted components (22).

Fig. 3 Position-velocity (p-v) diagrams for the [C ii] 158-μm emission.

(A and B) Distance is calculated from the center of the [C ii] 158-μm emission along the observed major axis of the galaxy (Fig. 1). For ALMA J081740.86+135138.2, a model of a simple uniformly rotating disk is shown in contours (22). The agreement between this model and the data suggests that the [C ii] 158-μm line originates from a cool gaseous disk. For ALMA J120110.26+211756.2,the [C ii] 158-μm emission appears more complicated, albeit at alower signal-to-noise ratio. mJy, millijansky.

Far-infrared dust-continuum emission cospatial with the [C ii] 158-μm emission was also detected in the ALMA observations (Fig. 1). The dust continuum yields estimates of the SFR, as emission is due to reprocessed starlight, mainly from the youngest stars (22). We find that both galaxies are forming stars at moderately high rates, with SFR of Embedded Image for ALMA J081740.86+135138.2 and Embedded Image for ALMA J120110.26+211756.2. The total far-infrared (TIR) luminosities of the galaxies are estimated by fitting a modified blackbody spectrum to the observed FIR dust-continuum measurement. The properties of the galaxies are tabulated in table S2.

The observations enable a comparison between the properties of these absorption-selected galaxies and the properties of emission-selected galaxies observed at similar redshifts, such as Lyman-break galaxies (LBGs). LBGs are star-forming gala xies identified in photometric studies by their characteristic rest-frame ultraviolet luminosity deficit. This deficit is due to the absorption of photons blueward of the Lyman continuum break (26). The observed SFRs of the two absorption-selected galaxies are comparable to the SFR of bright LBGs at similar redshift (27). However, by selecting the DLAs on their high metallicity, we are probing the massive end of the distribution of DLA host galaxies (11, 12). The typical galaxies associated with DLAs likely have a smaller mass, and thus a smaller SFR, consistent with nondetections reported in the literature (17).

Plotting the [C ii] 158-μm line luminosity versus the SFR for high-redshift (z ≥ 1) galaxies (Fig. 4A) shows that our absorption-selected galaxies occupy the same parameter space as high-redshift, emission-selected galaxies and fall within 1 SD of the correlation found in the local universe between these two observables (28). Similarly, the two absorption-selected galaxies fall within the same parameter space as their high-redshift emission-selected counterparts in a plot of the ratio of [C ii] 158-μm line luminosity to TIR luminosity versus TIR luminosity (Fig. 4B). This strengthens the assertion that high-metallicity, absorption-selected galaxies are similar to moderately star-forming, emission-selected galaxies, such as the massive end of the LBG population. This assertion is further corroborated by the agreement between our dynamical mass estimate in ALMA J081740.86+135138.2 and the dynamical mass estimates of comparable emission-selected galaxies at z ~ 5 (27). The observed connection between massive LBGs and high-metallicity DLAs supports results from earlier DLA studies at lower redshifts (9, 29) and cross-correlation studies with the Ly-α forest (30).

Fig. 4 Comparison of observed properties of the DLA hosts with local and high-redshift galaxies.

(A) Correlation between the [C ii] 158-μm luminosity and star-formation rate (SFR) for a sample of z > 1 galaxies categorized by galaxy type (23). Also plotted is the correlation for galaxies in the local universe (blue solid line), along with the associated scatter (blue dashed lines) (28). (B) Ratio of [C ii] 158-μm luminosity to total infrared (TIR) luminosity, LTIR, as a function of LTIR for the same sample of z > 1 galaxies. Plotted in gray are a sample of local galaxies and local (ultra)luminous infrared galaxies, (U)LIRGs. The high-metallicity, absorption-selected galaxies occupy the same parameter space as moderately star-forming, emission-selected galaxies at similar redshift, suggesting that they have similar characteristics. Embedded Image, luminosity of the Sun. Error bars indicate 1σ uncertainty.

We can use the impact parameter estimates to probe the physical extent of the H i gas around absorption-selected galaxies. The observed impact parameters correspond to physical (proper) distances of 42 and 18 kpc at the DLA redshifts for ALMA J081740.86+135138.2 and ALMA J120110.26+211756.2, respectively. These distances are substantially larger than the extent of the [C ii] 158-μm line emission, which extends to ~5 kpc, and indicate that these galaxies have large reservoirs of H i relatively far away from their star-forming regions. The observed distances also require that the emission and absorption lines arise in physically separated gas. Therefore, the absorbing gas must probe either H i gas associated with a satellite galaxy, an enriched neutral outflow from the galaxy, or H i gas in the inner circumgalactic medium/extended disk of the galaxy. The first explanation is disfavored, because the satellite galaxies would need to be metal-enriched and show highly turbulent velocity dispersions of several hundreds of kilometers per second, indicative of massive star-forming systems. However, little star formation is observed, as no [C ii] 158-μm emission is seen at the position of the absorber.

The absorbing gas is therefore more likely to reside in the inner gaseous halo of each DLA host galaxy. For the DLA toward SDSS J081740.52+135134.5, the gas is systematically blueshifted along the same direction and with the same magnitude as the rotation of the cool gas disk observed in [C ii] 158-μm line emission. This could indicate that the gas detected in absorption is co-rotating in an extended disk. Most simulations predict precisely such an extended planar configuration fed by cold flows (31), with properties similar to those observed in the z = 4.26 DLA (32). The DLA toward SDSS J120110.31+211758.5 is harder to classify because its absorption is seen over the full velocity range of the emission profile, which spans almost 500 km s–1. This is more indicative of a large-scale outflow or a highly perturbed system, which is corroborated by the spatial shift in the [C ii] 158-μm line emission (fig. S2). In both cases, it is clear that the DLA host galaxy has effectively enriched its inner gaseous halo, in agreement with recent simulations (33).

Our results indicate that the galaxies giving rise to high-metallicity DLAs have characteristics similar to the high-mass end of the LBG population (9, 30) and are embedded in a large reservoir of neutral H i gas. This gas is being enriched by the galaxy but is bound to it, as there is almost no systematic velocity offset between the metals seen in absorption and the [C ii] 158-μm line in emission. This observation suggests that this halo gas will eventually accrete back onto the galaxy, providing enriched gas for future star formation.

Supplementary Materials

Materials and Methods

Figs. S1 and S2

Tables S1 to S4

References (3445)

References and Notes

  1. Materials and methods are available as supplementary materials.
Acknowledgments: Support for this work was provided by the NSF through award SOSPA2-002 from the National Radio Astronomy Observatory (NRAO). M.N. and J.X.P. are partially supported by a grant from the NSF (AST-1412981). N.K. acknowledges support from the Department of Science and Technology via a Swarnajayanti Fellowship (DST/SJF/PSA-01/2012-13), and M.R. was partially supported by a NASA Postdoctoral Program fellowship. ALMA is a partnership of the European Southern Observatory (ESO) (representing its member states), NSF (USA), and the National Institute of Natural Sciences (Japan), together with the National Research Council (Canada), the National Science Council and Academia Sinica Institute of Astronomy and Astrophysics (Taiwan), and Korea Astronomy and Space Science Institute (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, Associated Universities Inc. (AUI)/NRAO, and National Astronomical Observatory of Japan. The NRAO is a facility of the NSF operated under cooperative agreement by AUI. Part of the data presented herein were obtained at the W. M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California, and NASA. The observatory was made possible by the generous financial support of the W. M. Keck Foundation. The data reported in this paper are available though the ALMA archive ( with project code ADS/JAO.ALMA#2015.1.01564.S and the Keck Observatory archive ( with program ID U163HR.

Stay Connected to Science

Navigate This Article