The nature and origin of low-redshift O VI absorbers

The O VI ion observed in quasar absorption-line spectra is the most accessible tracer of the cosmic metal distribution in the low-redshift (z < 0.5) intergalactic medium (IGM). We explore the nature and origin of O VI absorbers using cosmological hydrodynamic simulations including galactic outflows with a range of strengths. We consider the effects of ionization background variations, non-equilibrium ionization and cooling, uniform metallicity and small-scale (sub-resolution) turbulence. Our main results are as follows. (1) IGM O VI is predominantly photo-ionized with T approximate to 10(4.2 +/- 0.2) K. A key reason for this is that O VI absorbers preferentially trace overenriched (by similar to x 5) regions of the IGM at a given density, which enhances metal-line cooling such that absorbers can cool to photo-ionized temperatures within a Hubble time. As such, O VI is not a good tracer of the warm-hot intergalactic medium. (2) The predicted O VI properties fit observables if and only if sub-resolution turbulence is added, regardless of any other model variations. The required turbulence increases with O VI absorber strength. Stronger absorbers arise from more recent outflows, so qualitatively this can be understood if IGM turbulence dissipates on the order of a Hubble time. The amount of turbulence is consistent with other examples of turbulence observed in the IGM and galactic haloes. (3) Metals traced by O VI and HI do not trace exactly the same baryons, but reside in the same large-scale structure. Our simulations reproduce observed alignment statistics between O VI and HI, yet aligned absorbers typically have O VI arising from cooler gas, and for stronger absorbers lower densities, than HI. Owing to peculiar velocities dominating the line structure, coincident absorption often arises from spatially distinct gas. (4) Photo-ionized O VI traces gas in a variety of environments, and is not directly associated with the nearest galaxy, though is typically nearest to similar to 0.1L* galaxies. Weaker O VI components trace some of the oldest cosmic metals. (5) Very strong absorbers (EW greater than or similar to 100 m angstrom) are more likely to be collisionally ionized, tracing more recent enrichment (less than or similar to 2 Gyr) within or near galactic haloes.