Skip to main content
SearchLoginLogin or Signup

Tracing Million-Degree Gas in Galaxy Halos with Coronal Broad Ly α Absorbers

Predictions from a semi-analytic model confronted with observational data

Published onDec 01, 2022
Tracing Million-Degree Gas in Galaxy Halos with Coronal Broad Ly α Absorbers

Spiral galaxies like the Milky Way are believed to be surrounded by large amounts of diffuse gas that is gravitationally bound to the galaxy’s potential well and extends to its virial radius (and beyond): the circumgalactic medium (CGM). Among all the CGM gas phases, the shock-heated, hot (million-degree) phase is particularly difficult to observe, owing to the very low density of the gas and its high degree of ionization. X-ray emission from the hot coronal plasma of external galaxies has been studied using different instruments ([1][2][3]), indicating that the coronae of Milky Way-type galaxies typically contain ~ 1010-1011 M of gas, exceeding the baryonic mass contribution of the cooler CGM phases by almost two orders of magnitude (see also [4]). For the Milky Way, the X-ray lines of highly-ionized oxygen, O VII and O VIII, also represent important tracers of hot, circumgalactic gas, as they can be observed either in absorption against X-ray-bright AGNs or in emission (e.g., [5][6]).

In this Perspectives article, I discuss the possibility to use thermally broadened H I Lyα absorption lines as tracers for the hot CGM around galaxies. Hot coronal gas around Milky-Way like galaxies is expected to have temperatures close to the virial temperatures of their host halos, typically a few 106 K. At such high temperatures, the gas is almost fully ionized by collisions, with neutral hydrogen fractions <10-5. Photoionization from radiation originating in the UV background and in the host galaxies themselves is mostly relevant for cooler gas with T < 106 K ([7]).

Figure 1: Illustration of our modeling approach of CBLAs

Although the neutral gas fraction in such gas is tiny, there exist a sufficient number of neutral hydrogen atoms along a sightline that passes through the hot halo of a Milky Way-type galaxy (Fig. 1) to create a detectable Ly α absorption signal. The resulting so-called broad Lyα absorber (BLA) is shallow and broad, owing to the substantial thermal line broadening caused by the high gas temperature. In the following, I denote BLAs that trace the hot coronal gas of galaxies as coronal broad Lyα absorbers or CBLAs. In anticipation of the modeling results, Figure 2 shows an example of the spectral appearance of a CBLA that passes through the hot halo of an L* galaxy at an impact parameter of D=100 kpc.

Figure 2: Example of a synthetic CBLA line that traces hot (million degree) coronal gas in the halo of an L* galaxy along a sightline with an impact parameter of D = 100 kpc (with an HI column density of log N(HI) = 12.9 and a Doppler parameter of b = 105 km s−1; upper panel: no noise, lower panel: S/N=50 per pixel)

Broad Lyα absorbers (i.e., “normal” BLAs) have been previously studied by us and other research groups to trace the missing baryons in the warm–hot IGM (WHIM; [8][9][10]). BLAs that possibly are associated with warm–hot gas in the halos of individual galaxies (i.e., CBLA candidates) have been reported regularly in previous studies (see, e.g., [11]), but a systematic investigation on how such broad absorbers that are related to the hot coronal gas around their host galaxies was not presented until 2020, when our pilot study was published during the Coronavirus pandemic [12]. The major advantage of analyzing circumgalactic CBLAs compared to intergalactic BLAs are that (i) we know exactly where we should look for them, namely along sightlines that pass galaxies within their virial radii at radial velocities defined by these galaxies, (ii) the hot gas is confined in a much smaller volume (i.e., within the virial radius of the galaxies), eliminating large-scale gas flows as a line-broadening mechanism, and (iii) the temperature (and thus the ionization fraction) of the collisionally ionized gas is expected to scale with a galaxy’s virial mass, allowing us to predict the CBLA absorption properties for each individual galaxy/sightline pair.

However, the CGM is multi-phase, and therefore the cooler (less ionized) gas phases dominate the H I optical depth in most CGM absorbers. As a result, most CBLAs are expected to be embedded in (or hidden by) complex, multi-component H I Lyα absorption systems. This is demonstrated in Figure 3, which shows CBLAs at z=0.005 and z=0.139 in the HST/COS and HST/STIS spectra of 3C273 and PG 1116+215.

Figure 3: CBLA candidate systems in the COS and STIS spectra of the AGN 3C273 and PG 1116+215. The left panels show the raw spectral data in the overall wavelength ranges where CBLA absorption in the halos of intervening galaxies is expected. Galaxy data and impact parameters are listed in the panels. The gray shaded areas indicate the expected range for CBLA absorption based on the accuracy of the galaxy redshifts and allowing for co-rotation of the coronal gas with the disk. Theright panels show the data together with the best-fitting multi-component models of the Lyα absorbers (black solid line) and the modeled CBLA absorption (red shaded area)

The expected strength and shape of a CBLA that is associated with a given galaxy can be analytically modeled as a function of galaxy halo mass and galaxy impact parameter assuming that the hot gas is confined in the galaxy’s DM halo in quasi-hydrostatic equilibrium. For this, I have used in [12] the formalism developed by Maller & Bullock [13] and adjusted/extended it to predict CBLA aborption profiles. First, I derived analytic equations for the radial density and temperature profiles of the residual hydrostatic hot gas halo in a NFW potential for a given halo mass assuming gas cooling and fragmentation under realistic conditions. I then calculated radial neutral hydrogen density profiles assuming collisional ionization equilibrium, for which the neutral gas fraction depends only on the local coronal gas temperature (which is close to the halo’s virial temperature). From this I derived the neutral hydrogen column density of a CBLA, N(HI), along a halo sightline at impact parameter D by integrating the neutral gas density the along the path through the coronal gas distribution. All individual steps are explained in detail in [12].

I have applied this methodology to the local galaxy population to investigate the statistical properties of the CBLAs in the local Universe. For this, I generated 2134 sightlines passing through 11 model halos at z=0 in the mass range log (MV/M) = 10.6–12.6 (corresponding to galaxy luminosities in the range L/L*=0.1-10.0) at impact parameters D<Rvir (in steps of 1 kpc). In the resulting CBLA mock sample, 84 percent of the absorbers have weak logarithmic HI column densities in the range log N(HI)=12.4–13.4. The column-density distribution (Fig. 4, upper panel) peaks at log N(H I)≈12.9, a value that can be regarded as “characteristic” for CBLAs in the model set. The distribution of HI Doppler parameters/b-values peaks at a characteristic value of b≈140 km s−1, where 82 percent of the modelled CBLAs have b-values in the range b=70–200 km s−1 (Fig. 3, lower panel).

The distributions shown in Figure 3 indicate that CBLAs span a broad range in N(H I) and b. However, detecting broad, shallow absorption features in UV data with limited S/N is challenging. The detection significance depends on both the depth and the width of the absorption as well as on the local S/N.

The cosmological cross section of CBLAs (here expressed as number of absorbers per unit redshift) depends on both the space density of galaxies (as can be determined from the galaxy luminosity function) and the projected geometrical cross section of the coronal gas in each halo (out to the virial radius), which can be inferred directly from our model. Doing this for the entire CBLA mock sample and considering different signal-to-noise (S/N) levels in the spectra, we obtain total number densities per unit redshift of dN/dz(CBLA) = 2.6 for idealized, noise-free spectra, dN/dz(CBLA) = 2.1 for S/N=100, and dN/dz(CBLA) = 0.8 for S/N=50 among the halos in the adopted mass range. These numbers demonstrate that CBLAs are expected to have a substantial cosmological absorption cross section.

Because the strongest CBLAs sample the inner regions of galaxy halos, CBLAs in real UV absorption spectra are expected to blend with narrow (and predominantly stronger) H I features stemming from the warm–cool (T<105 K) CGM that traces infalling and outflowing gaseous material. Therefore, a fair fraction of CBLAs in extragalactic UV spectra are expected to be hidden in multi-component H I Lyα profiles and may not be readily visible.

To test these predictions on the local CBLA population, our group in Potsdam is currently searching for CBLA candidate systems in all available HST/COS and HST/STIS archival QSO data starting with data sets that we have used in previous absorption-line studies [14][15][16]. For such an observational survey, we are particularly interested in QSO sightlines that are known to pass the halos of nearby massive galaxies within their virial radii and for which UV spectral data with good S/N are available. Two examples for CBLAs along the sightlines towards 3C273 and PG 1116+215 are shown in Figure 3, wheras two additional examples are presented in our pilot study ([12]).

As can be seen, these Ly α systems all are dominated by a strong absorption component that traces cooler gas components in the inner halos of these galaxies. However, the inclusion of a broad, shallow absorption component (i.e., a CBLA) in the absorber models is required to account for the observed flux depressions in the wings of the strong Lyα absorption and to provide an optimum fit to the COS and STIS spectral data. As mentioned earlier, also other authors have identified and discussed broad Ly α absorption as potential tracers for warm–hot circumgalactic gas in their analyses [11][17][18][19]. Yet, a systematic investigation of these features with regard to the expected spectral signatures of shock-heated coronal gas has not been provided so far.

The semi-analytic model presented in [12] and discussed here provides the theoretical basis for a systematic study of such interesting systems. A CBLA survey (based on the entire HST/COS archival data base) is currently prepared by us and will be presented in a forthcoming paper.

No comments here
Why not start the discussion?