Galaxies are thought to be fed by the continuous accretion of intergalactic gas, but direct observational evidence has been elusive. The accreted gas is expected to orbit about the galaxys halo, delivering not just fuel for star-formation but also an
gular momentum to the galaxy, leading to distinct kinematic signatures. Here we report observations showing these distinct signatures near a typical distant star-forming galaxy where the gas is detected using a background quasar passing 26 kpc from the host. Our observations indicate that gas accretion plays a major role in galaxy growth since the estimated accretion rate is comparable to the star-formation rate.
It is well established that MgII absorption lines detected in background quasar spectra arise from gas structures associated with foreground galaxies. The degree to which galaxy evolution is driven by the gas cycling through halos is highly uncertain
because their gas mass density is poorly constrained. Fitting the MgII equivalent width (W) distribution with a Schechter function and applying the N(HI)-W correlation of Menard & Chelouche, we computed Omega(HI)_MgII ~ Omega(HI)_halo =(1.41 +0.75 -0.44)x10^-4 for 0.4<z<1.4. We exclude DLAs from our calculations so that Omega(HI)_halo comprises accreting and/or outflowing halo gas not locked up in cold neutral clouds. We deduce the cosmic HI gas mass density fraction in galactic halos traced by MgII absorption is Omega(HI)_halo/Omega(HI)_DLA=15% and Omega(HI)_halo/Omega_b=0.3%. Citing several lines of evidence, we propose infall/accretion material is sampled by small W whereas outflow/winds are sampled by large W, and find Omega(HI)_infall is consistent with Omega(HI)_outflow for bifurcation at W=1.23^{+0.15}_{-0.28}AA; cold accretion would then comprise no more than ~7% of of the total HI mass density. We discuss evidence that (1) the total HI mass cycling through halos remains fairly constant with cosmic time and that the accretion of HI gas sustains galaxy winds, and (2) evolution in the cosmic star formation rate depends primarily on the rate at which cool HI gas cycles through halos.
We have used GIM2D to quantify the morphological properties of 40 intermediate redshift MgII absorption-selected galaxies (0.03<Wr(2796)<2.9 Ang), imaged with WFPC-2/HST, and compared them to the halo gas properties measured form HIRES/Keck and UVES/
VLT quasar spectra. We find that as the quasar-galaxy separation, D, increases the MgII equivalent decreases with large scatter, implying that D is not the only physical parameter affecting the distribution and quantity of halo gas. Our main result shows that inclination correlates with MgII absorption properties after normalizing out the relationship (and scatter) between the absorption properties and D. We find a 4.3 sigma correlation between Wr(2796) and galaxy inclination, normalized by impact parameter, i/D. Other measures of absorption optical depth also correlate with i/D at greater than 3.2 sigma significance. Overall, this result suggests that MgII gas has a co-planer geometry, not necessarily disk-like, that is coupled to the galaxy inclination. It is plausible that the absorbing gas arises from tidal streams, satellites, filaments, etc., which tend to have somewhat co-planer distributions. This result does not support a picture in which MgII absorbers with Wr(2796)<1A are predominantly produced by star-formation driven winds. We further find that; (1) MgII host galaxies have quantitatively similar bulge and disk scale length distribution to field galaxies at similar redshifts and have a mean disk and bulge scale length of 3.8kpc and 2.5kpc, respectively; (2) Galaxy color and luminosity do not correlate strongly with absorption properties, implying a lack of a connection between host galaxy star formation rates and absorption strength; (3) Parameters such as scale lengths and bulge-to-total ratios do not significantly correlate with the absorption parameters, suggesting that the absorption is independent of galaxy size or mass.
We have directly compared MgII halo gas kinematics to the rotation velocities derived from emission/absorption lines of the associated host galaxies. Our 0.096<z<0.148 volume-limited sample comprises 13 ~L* galaxies, with impact parameters of 12-90 k
pc from background quasars sight-lines, associated with 11 MgII absorption systems with MgII equivalent widths 0.3< W_r(2796)<2.3A. For only 5/13 galaxies, the absorption resides to one side of the galaxy systemic velocity and trends to align with one side of the galaxy rotation curve. The remainder have absorption that spans both sides of the galaxy systemic velocity. These results differ from those at z~0.5, where 74% of the galaxies have absorption residing to one side of the galaxy systemic velocity. For all the z~0.1 systems, simple extended disk-like rotation models fail to reproduce the full MgII velocity spread, implying other dynamical processes contribute to the MgII kinematics. In fact 55% of the galaxies are counter-rotating with respect to the bulk of the MgII absorption. These MgII host-galaxies are isolated, have low star formation rates (SFRs) in their central regions (<1 Msun/yr), and SFRs per unit area well below those measured for galaxies with strong winds. The galaxy NaID (stellar+ISM) and MgIb (stellar) absorption line ratios are consistent with a predominately stellar origin, implying kinematically quiescent interstellar media. These facts suggest that the kinematics of the MgII absorption halos for our sample of galaxies are not influenced by galaxy--galaxy environmental effects, nor by winds intrinsic to the host galaxies. For these low redshift galaxies, we favor a scenario in which infalling gas accretion provides a gas reservoir for low-to-moderate star formation rates and disk/halo processes.
(Abridged) We performed a spectroscopic galaxy survey, complete to m<20.3 (L_B>0.15L_B* at z=0.3), within 100x100 of the quasar Q1127-145 (z=1.18). The VLT/UVES quasar spectrum contains three z<0.33 MgII absorption systems. We obtained eight new gala
xy redshifts, adding to the four previously known, and galaxy star formation rates and metallicities were computed where possible. A strong MgII system [W_r(2796)=1.8A], which is a known DLA, had three previously identified galaxies; we found two additional galaxies associated with this system. These five galaxies form a group with diverse properties, such as a luminosity range of 0.04<L_B<0.63L_B*, an impact parameter range of 17<D<241kpc and velocity dispersion of 115km/s. The DLA group galaxy redshifts span beyond the 350km/s velocity spread of the metallic absorption lines of the DLA itself. The two brightest group galaxies have SFRs of a few Msun/yr and should not have strong winds. We have sufficient spectroscopic information to directly compare three of the five group galaxies (emission-line) metallicities with the DLA (absorption) metallicity: the DLA metallicity is 1/10th solar, substantially lower than the three galaxies which range between less than 1/2 solar to solar metallicity. HST/WFPC-2 imaging shows perturbed morphologies for the three brightest group galaxies, with tidal tails extending 25kpc. We favor a scenario where the DLA absorption originates from tidal debris in the group environment. Another absorber exhibits weak MgII absorption [W_r(2796)=0.03A] and had a previously identified galaxy at a similar redshift. We have identified a second galaxy associated with this system. Both galaxies have solar metallicities and unperturbed morphologies. The SFR of one galaxy is much lower than expected for strong outflows. Finally, we have identified five galaxies at large impact parameters with no associated MgII absorption.
We obtained ESI/Keck rotation curves of 10 MgII absorption selected galaxies (0.3 < z < 1.0) for which we have WFPC-2/HST images and high resolution HIRES/Keck and UVES/VLT quasar spectra of the MgII absorption profiles. We perform a kinematic compar
ison of these galaxies and their associated halo MgII absorption. For all 10 galaxies, the majority of the absorption velocities lie in the range of the observed galaxy rotation velocities. In 7/10 cases, the absorption velocities reside fully to one side of the galaxy systemic velocity and usually align with one arm of the rotation curve. In all cases, a constant rotating thick-disk model poorly reproduces the full spread of observed MgII absorption velocities when reasonably realistic parameters are employed. In 2/10 cases, the galaxy kinematics, star formation surface densities, and absorption kinematics have a resemblance to those of high redshift galaxies showing strong outflows. We find that MgII absorption velocity spread and optical depth distribution may be dependent on galaxy inclination. To further aid in the spatial-kinematic relationships of the data, we apply quasar absorption line techniques to a galaxy (v_c=180 km/s) embedded in LCDM simulations. In the simulations, MgII absorption selects metal enriched halo gas out to roughly 100 kpc from the galaxy, tidal streams, filaments, and small satellite galaxies. Within the limitations inherent in the simulations, the majority of the simulated MgII absorption arises in the filaments and tidal streams and is infalling towards the galaxy with velocities between -200 < v_r < -180 km/s. The MgII absorption velocity offset distribution (relative to the simulated galaxy) spans ~200 km/s with the lowest frequency of detecting MgII at the galaxy systematic velocity.
We examine halo gas cross sections and covering fractions, f_c, of intermediate redshift MgII absorption selected galaxies. We computed statistical absorber halo radii, R_x, using current values of dN/dz and Schechter luminosity function parameters,
and have compared these values to the distribution of impact parameters and luminosities from a sample of 37 galaxies. For equivalent widths W_r(2796) > 0.3 Ang, we find 43 < R_x < 88 kpc, depending on the lower luminosity cutoff and the slope, beta, of the Holmberg-like luminosity scaling, R propto L^beta. The observed distribution of impact parameters, D, are such that several absorbing galaxies lie at D > R_x and several non-absorbing galaxies lie at D < R_x. We deduced f_c must be less than unity and obtain a mean of <f_c> ~ 0.5 for our sample. Moreover, the data suggest halo radii of MgII absorbing galaxies do not follow a luminosity scaling with beta in the range of 0.2-0.28, if f_c= 1 as previously reported. However, provided f_c~0.5, we find that halo radii can remain consistent with a Holmberg-like luminosity relation with beta ~ 0.2 and R* = R_x/sqrt(f_c)= 110 kpc. No luminosity scaling (beta=0) is also consistent with the observed distribution of impact parameters if f_c < 0.37. The data support a scenario in which gaseous halos are patchy and likely have non-symmetric geometric distributions about the galaxies. We suggest halo gas distributions may not be govern primarily by galaxy mass/luminosity but also by stochastic processes local to the galaxy.
We have explored the galaxy disk/extended halo gas kinematic relationship using rotation curves (Keck/ESI) of ten intermediate redshift galaxies which were selected by MgII halo gas absorption observed in quasar spectra. Previous results of six edge-
on galaxies, probed along their major axis, suggest that observed halo gas velocities are consistent with extended disk-like halo rotation at galactocentric distances of 25-72 kpc. Using our new sample, we demonstrate that the gas velocities are by and large not consistent with being directly coupled to the galaxy kinematics. Thus, mechanisms other than co-rotation dynamics (i.e., gas inflow, feedback, galaxy-galaxy interactions, etc.) must be invoked to account for the overall observed kinematics of the halo gas. In order to better understand the dynamic interaction of the galaxy/halo/cosmic web environment, we performed similar mock observations of galaxies and gaseous halos in Lambda-CDM cosmological simulations. We discuss an example case of a z=0.92 galaxy with various orientations probing halo gas at a range of positions. The gas dynamics inferred using simulated quasar absorption lines are consistent with observational data.
