We demonstrate the presence of an extended and massive circumgalactic medium (CGM) around Messier 31 using archival HST COS ultraviolet spectroscopy of 18 QSOs projected within two virial radii of M31 (Rvir=300 kpc). We detect absorption from SiIII a
t -300<vLSR}<-150 km/s toward all 3 sightlines at R<0.2Rvir, 3 of 4 sightlines at 0.8<R/Rvir<1.1, and possibly 1 of 11 at 1.1<R/Rvir<1.8. We present several arguments that the gas at these velocities observed in these directions originates from the CGM of M31 rather than the Local Group or Milky Way CGM or Magellanic Stream. We show that the dwarf galaxies located in the CGM of M31 have very similar velocities over similar projected distances from M31. We find a non-trivial relationship only at these velocities between the column densities (N) of all the ions and R, whereby N decreases with increasing R. Singly ionized species are only detected in the inner CGM of M31 at R<0.2Rvir. At R<0.8 Rvir, the covering fraction is close to unity for SiIII and CIV (fc~60%-97% at the 90% confidence level), but drops to fc<10-20% at R>Rvir. We show that the M31 CGM gas is bound, multiphase, predominantly ionized (i.e., HII>>HI), and becomes more highly ionized gas at larger R. We estimate using SiII, SiIII, and SiIV a CGM metal mass of at least 2x10^6 Msun and gas mass of >3x10^9(Zsun/Z) Msun within 0.2 Rvir, and possibly a factor ~10 larger within Rvir, implying substantial metal and gas masses in the CGM of M31. Compared with galaxies in the COS-Halos survey, the CGM of M31 appears to be quite typical for a L* galaxy.
We present UV and optical observations from the Cosmic Origins Spectrograph on the Hubble Space Telescope and Keck of a z= 0.27395 Lyman limit system (LLS) seen in absorption against the QSO PG1630+377. We detect H I absorption with log N(HI)=17.06pm
0.05 as well as Mg II, C III, Si III, and O VI in this system. The column densities are readily explained if this is a multi-phase system, with the intermediate and low ions arising in a very low metallicity ([Mg/ H] =-1.71 pm 0.06) photoionized gas. We identify via Keck spectroscopy and Large Binocular Telescope imaging a 0.3 L_* star-forming galaxy projected 37 kpc from the QSO at nearly identical redshift (z=0.27406, Delta v = -26 kms) with near solar metallicity ([O/ H]=-0.20 pm 0.15). The presence of very low metallicity gas in the proximity of a near-solar metallicity, sub-L_* galaxy strongly suggests that the LLS probes gas infalling onto the galaxy. A search of the literature reveals that such low metallicity LLSs are not uncommon. We found that 50% (4/8) of the well-studied z < 1 LLSs have metallicities similar to the present system and show sub-L_* galaxies with rho < 100 kpc in those fields where redshifts have been surveyed. We argue that the properties of these primitive LLSs and their host galaxies are consistent with those of cold mode accretion streams seen in galaxy simulations.
We present a survey for optically thick Lyman limit absorbers at z<2.6 using archival Hubble Space Telescope observations with the Faint Object Spectrograph and Space Telescope Imaging Spectrograph. We identify 206 Lyman limit systems (LLSs) increasi
ng the number of catalogued LLSs at z<2.6 by a factor of ~10. We compile a statistical sample of 50 tau_LLS > 2 LLSs drawn from 249 QSO sight lines that avoid known targeting biases. The incidence of such LLSs per unit redshift, l(z)=dn/dz, at these redshifts is well described by a single power law, l(z) = C1 (1+z)^gamma, with gamma=1.33 +/- 0.61 at z<2.6, or with gamma=1.83 +/- 0.21 over the redshift range 0.2 < z < 4.9. The incidence of LLSs per absorption distance, l(X), decreases by a factor of ~1.5 over the ~0.6 Gyr from z=4.9 to 3.5; l(X) evolves much more slowly at low redshifts, decreasing by a similar factor over the ~8 Gyr from z=2.6 to 0.25. We show that the column density distribution function, f(N(HI)), at low redshift is not well fitted by a single power law index (f(N(HI)) = C2 N(HI)^(-beta)) over the column density range 13 < log N(HI) < 22 or log N(HI) >17.2. While low and high redshift f(N(HI)) distributions are consistent for log N(HI)>19.0, there is some evidence that f(N(HI)) evolves with z for log N(HI) < 17.7, possibly due to the evolution of the UV background and galactic feedback. Assuming LLSs are associated with individual galaxies, we show that the physical cross section of the optically thick envelopes of galaxies decreased by a factor of ~9 from z~5 to 2 and has remained relatively constant since that time. We argue that a significant fraction of the observed population of LLSs arises in the circumgalactic gas of sub-L* galaxies.
