4. GLOBAL HOT GAS IN AND AROUND OTHER GALAXIES

The X-ray absorption line spectroscopy can also be used to probe the hot CGM around other galaxies. Yao et al. have recently studied the hot CGM along the sight lines toward luminous AGNs [24]. As an example, Fig. 2 shows a Chandra grating spectrum of PKS 2155-304. No absorption line is apparent at the redshifts of the three groups of galaxies with the impact distances smaller than 500 kpc of the sight line (in particular, the group at the highest redshift contains two galaxies brighter than L_*). The same is true when the spectrum is folded according to the redshifts of the individual groups. To increase the counting statistics further, Yao et al. have stacked spectra from the Chandra grating observations of eight AGNs with such intervening galaxies or groups. No significant absorption is detected for any of these individual systems or in the final stacked spectrum with a total equivalent exposure of about 10 Ms! Upper limits to the mean column densities of various ion species per galaxy or group are then estimated. In particular, they find that N_OVII 6 × 10¹⁴ cm^-2, consistent with the constraints for the CGM around our Galaxy. They have estimated the total mass contained in the CGM as M_CGM 0.6 × (0.5 / f_OVII) × (0.3A / A) × (R / 500 kpc)² × 10¹¹ M, where f_OVII, A, and R are the ionization fraction of O VII, metal abundance, and the radius of the hot CGM, respectively. This is in contrast to the expected baryon mass 2 × 10¹¹ M for the halo of a Milky Way-type galaxy or a typical galaxy group [25]. Thus the bulk of the CGM unlikely resides in such a chemically enriched warm-hot phase at temperatures ranging from 10^5.5-10^6.5 K (Fig. 1), which our X-ray absorption line spectroscopy is sensitive to. This conclusion has strong implications for understanding the accumulated effect of the stellar and AGN feedback on the galactic ecosystem (see the discussion section).

Figure 2. Part of the Chandra grating spectrum of PKS 2155-304 with a total exposure of 1 Ms. Various absorption lines (marked in blue) are detected and are produced by hot gas in and around the Galaxy (zero redshift). The red-shaded regions enclose spectral ranges of the same lines for three red-shifted groups of galaxies located within the 500 kpc impact distance of the AGN sight line.

To study the effect of ongoing stellar and AGN feedback, one can map out diffuse X-ray emission from hot gas in and around nearby galaxies of various masses and star formation rates. Much attention has been paid to the feedback in starburst and massive elliptical galaxies, which are relatively bright in diffuse X-ray emission. Chandra observations have shown convincingly that the AGN feedback is important in shaping the morphology and thermal evolution of hot gas in massive elliptical galaxies, particularly those at centers of galaxy groups and clusters ([26] and references therein). The asymmetry in the global diffuse X-ray morphology is correlated with radio and X-ray luminosities of AGNs in elliptical galaxies, even in rather X-ray-faint ones [27]. This calls into question the hydrostatic assumption commonly used in order to infer the gravitational mass distribution in such galaxies. Nevertheless, the hydrostatic assumption may hold approximately for hot gas around the central supermassive black holes (SMBHs), if they are in a sufficiently quiescent state. The SMBH masses may then be measured from spatially resolved X-ray spectroscopy of the hot gas. Humphrey et al. have made such mass measurements for four SMBHs with Chandra data [28]. Three of them already have mass determinations from the kinematics of either stars or a central gas disk. It is encouraging to find a good agreement between the measurements using the different methods. From this agreement, they further infer that no more than ~ 10%-20% of the ISM pressure around the SMBHs should be nonthermal.

The feedback in nuclear starburst galaxies is manifested in the so-called galactic superwinds driven by the mechanical energy injection from fast stellar winds and supernovae (SNe) of massive stars (e.g., [29, 30, 31]). The observed soft X-ray emission from a superwind typically has an elongated morphology along the minor axis of such a galaxy and is correlated well with extraplanar H-emitting features. This indicates that the detected hot gas arises primarily from the interaction between the superwind and cool gas. The superwind itself, believed to be very hot and low in density, is much difficult to detect. From a detailed comparison between Chandra data and hydrodynamic simulations, Strickland & Heckman infer that the superwind of M82 has a mean temperature of 3-8 × 10⁷ K and a mass outflowing rate of ~ 2 M yr^-1 [31]. Such energetic superwinds with little radiative energy loss must have profound effects on the large-scale CGM (e.g., [30]).

Recent X-ray observations have further shown the importance of the feedback in understanding even "normal" intermediate-mass galaxies (similar to the Milky Way and M31; e.g., [29, 30, 32, 33, 34, 35, 36, 37, 38, 39, 40]). Chandra, in particular, has unambiguously detected diffuse hot gas in and around normal disk galaxies. The total X-ray luminosity of the gas is well correlated with the star formation rate for such galaxies. The diffuse soft X-ray emission is shown to be strongly enhanced in recent star forming regions or spiral arms within an individual galaxy viewed face-on and is only slightly more diffuse than H emission (e.g., [33, 34]). This narrow appearance of spiral arms in X-ray conflicts the expectation from population synthesis models: the mechanical energy output rate from SNe should be nearly constant over a time period that is up to 10 times longer than the lifetime of massive ionizing stars. This means that SNe in inter-arm regions, where ISM density is generally low, must produce less soft X-ray emission than those in the arms. When disk galaxies are observed in an inclined angle (e.g., Fig. 3), the soft X-ray emission tends to appear as plumes, most likely representing blown-out hot gas heated in recent massive star forming regions and galactic spheroids. The cleanest perspectives of the extraplanar hot gas are obtained from the observations of edge-on disk galaxies such as NGC 891 [29] and NGC 5775 [38]. The observed diffuse X-ray emission typically does not extend significantly more than a few kpc away from the galactic disks. A claimed detection of the emission around the edge-on spiral NGC 5746 on larger scales was later proved to be due to an instrumental artifact [41]. Complementary observations from XMM-Newton and Suzaku give consistent results and provide improved spectral information on the hot gas (e.g., [35, 36, 40, 42, 43]). The morphology of the X-ray emission as well as its correlation with the star formation rate clearly shows that the extraplanar hot gas is primarily heated by the stellar feedback. In fact, the cooling of the hot gas accounts for only a small fraction of the expected energy input from massive stars alone (typically no more than a few %).

Figure 3. Chandra diffuse 0.5-1.5 keV intensity contours, overlaid on a Spitzer mid-IR image of the Sb galaxy NGC 2841. The one-sided morphology of the diffuse X-ray emission apparently represents outflows of hot gas from the tilted galactic disk with its northeast edge closer to us. The emission from the back side is largely absorbed by the cool gas in the disk. The crosses mark the positions of excised discrete X-ray sources.

The best example of mapping out hot gas in a galactic stellar bulge (or spheroid) is the discovery of an apparent hot gas outflow from the M31 bulge [37, 39]. This outflow with a 0.5-2 keV luminosity of ~ 2 × 10³⁸ erg s^-1 is driven apparently by the feedback from evolved stars in form of stellar mass loss and (Type) Ia SNe, because there is no evidence for an AGN or recent massive star formation in the bulge. The bipolar morphology of the truly diffuse soft X-ray emission further indicates that the outflow is probably influenced by the presence of strong vertical magnetic field, same as that observed in the central region of our Galaxy. The spectrum of the emission can be characterized by an optically-thin thermal plasma with a temperature of ~ 3 × 10⁶ K [37, 39, 44]. Similar analyses have been carried out so far only for a couple of other early-type galaxies (NGC 3379, [45]; NGC 5866, [46]). The faint stellar contribution, which has a significantly different spectral shape from LMXBs', still needs to be carefully accounted for in the analysis of other galaxies. In any case, it is clear that the diffuse X-ray luminosity accounts for at most a few % of the energy input from Ia SNe alone in a normal early-type galaxy. Where does the missing feedback go?

Another remarkable puzzle about the diffuse hot gas in elliptical galaxies is the apparent low metal abundances. As inferred from X-ray spectral fits, the abundances are typically sub-solar for low- and intermediate-mass galaxies with log(L_x) 41 to about solar for more massive ones (e.g., [47, 48]), substantially less than what are expected from the Ia SN enrichment (see the discussion section). Furthermore, a number of galaxies show a significant iron abundance drop toward their central regions, where the stellar feedback should be the strongest. Examples of this abundance drop include M87 [49], NGC 4472, NGC 5846 [50], and NGC 5044 [51]. The drop and the low abundance, if intrinsic, would then indicate that only a small portion of metals produced by Ia SNe is observed; the rest is either expelled or in a state that the present X-ray data are not sensitive to.

In summary, the following are the key discoveries made with Chandra observations of diffuse hot gas in and around normal galaxies:

• The hot gas generally consists of two components: 1) the disk component, which is sensitive to the star formation rate per unit area (e.g., [30, 38, 52, 43]); 2) the bulge component, the amount of which is proportional to the stellar mass. Significant diffuse soft X-ray emission is detected only within ~ 10 kpc of the disk and bulge of a galaxy.

• The gas typically has a mean characteristic temperature of ~ 10^6.3 K, although there is evidence for a considerably high temperature component (up to 10⁷ K) in and around galactic disks with active star formation [34, 38]. This high-temperature component may represent a significant galactic hot gaseous outflow. The interaction of this outflow with cool gas (via shocks, dynamic mixing, and/or charge exchanges) is responsible for much of the low-temperature one (e.g., [31, 38]).

• The cooling rate, as inferred from the diffuse X-ray emission, is insignificant, compared to the expected stellar feedback alone. The bulk of the feedback is missing, especially in inter-arm regions and in galactic spheroids.

• The hot CGM seems to have a very low density and may not account for the missing baryon matter expected in individual galactic dark matter halos.

While the above results are generally consistent with those found for our own Galaxy, the studies of nearby galaxies also show intriguing sub-structures in the diffuse hot gas distributions (e.g., the apparent positive radial temperature and metal abundance gradients in elliptical galaxies).