The warm ionized medium in spiral galaxies

This article reviews observations and models of the diffuse ionized gas that permeates the disk and halo of our Galaxy and others. It was inspired by a series of invited talks presented during an afternoon scientific session of the 65th birthday celebration for Professor Carl Heiles held at Arecibo Observatory in August 2004. This review is in recognition of Carl’s long-standing interest in and advocacy for studies of the ionized as well as the neutral components of the interstellar medium.

Figure 1

Composite all-sky velocity-integrated $\mathrm{H}\alpha$ map. Data from VTSS, SHASSA, and the WHAM northern sky survey (WHAM-NSS) have been combined to produce these very deep (EM $\gtrsim 1{\mathrm{cm}}^{-6}\mathrm{pc}$) emission maps. The two Aitoff-Hammer projections are centered at (top) $\mathcal{l}=0°$ and (bottom) $\mathcal{l}=180°$. Areas covered by the imaging surveys have arc-minute resolution while those only surveyed by WHAM have one-degree resolution. Emission is predominantly from the Galaxy, although the imaging surveys can contain other bright, extended sources from the local universe ($\mid v_{\mathrm{LSR}}\mid \lesssim 500\mathrm{km}{\mathrm{s}}^{-1}$; LMC, SMC, M31, etc.). Adapted from 45.

Figure 2

Diagnostic line ratio diagrams. A large portion of the Galaxy in the direction of the Perseus arm ($\mathcal{l}=130°$ to 160° and $b=-30°$ to $+30°$, approximately) has been surveyed in $\mathrm{H}\alpha$, [S II], and [N II] with WHAM. These panels show (left) the relationship between the ratios vs the intensity of $\mathrm{H}\alpha$ as well as (right) the relationship between the two ratios for the “local” gas component $(\mid v_{\mathrm{LSR}}\mid <15\mathrm{km}{\mathrm{s}}^{-1})$. A few specific spatial regions are highlighted with grayscale fill to show the effects of local ionizing sources: the planetary nebula (PN) S216, the H II region surrounding the $\mathrm{B}0.5+\mathrm{sdO}$ system $\varphi$ Per, and regions near O-star H II regions. Other data points sample the WIM. In the diagram on the right, the vertical dashed lines represent $T_e=5000$, 6000, 7000, 8000, 9000, and $10000\mathrm{K}$, from left to right. The slanted solid lines represent ${\mathrm{S}}^{+}∕\mathrm{S}=0.25$, 0.50, 0.75, and 1.00, lowest to highest slope. These WIM data reveal significant variations in $T_e$ and ${\mathrm{S}}^{+}∕\mathrm{S}$ from one line of sight to the next. In contrast, classical H II regions all cluster in the lower left corner of this diagram near $\left[\mathrm{N}\mathrm{II}\right]∕\mathrm{H}\alpha \approx 0.25$, $\left[\mathrm{S}\mathrm{II}\right]∕\mathrm{H}\alpha \approx 0.1$. From 79.

Figure 3

Elevated temperature in the WIM. Select directions toward (◼) brighter diffuse ionized regions show elevated line ratios in both $\left[\mathrm{N}\mathrm{II}\right]∕\mathrm{H}\alpha$ and $\left[\mathrm{N}\mathrm{II}\right]\lambda 5755∕\lambda 6584$ compared to (○) H II regions. Dashed lines mark select temperatures derived from each line ratio while the diagonal line traces “unity” in this derived temperature space. WIM directions have higher derived temperatures using either line ratio, but are also systematically above the line of unity, suggesting that measurable temperature inhomogeneities exist along these lines of sight. From 79.

Figure 4

Temperature $T_i$ and the mode $v_{\mathrm{NT}}$ (most probable value) of the nonthermal speed distribution of the diffuse H II region surrounding $\zeta$ Oph. The primary axes show the measured widths and errors of $\mathrm{H}\alpha$ and [S II] from 126 one-degree pointings obtained by WHAM within a $12\text{−}\text{degree}$ diameter region around $\zeta$ Oph. A grid of temperatures and nonthermal velocities has been superimposed to show the distribution of $T_i$ and $v_{\mathrm{NT}}$ within the H II region. The apparent scatter in these values is spatially correlated (see text). The line at the upper left is the line of equal widths.

Figure 5

The dependence of DIG detections in halos of galaxies on the normalized star formation rate (per unit area). The ratio of the FIR fluxes at 60 and $100\mu \mathrm{m}$ $(S_{60}∕S_{100})$ is given vs the FIR luminosity normalized to the optical galaxy diameter at 25th mag in units of ${10}^{40}\mathrm{erg}{\mathrm{s}}^{-1}{\mathrm{kpc}}^{-2}$. Filled symbols indicate detections of ${\mathrm{H}}^{+}$ gas in the halo while open symbols represent nondetections. Galaxies from the (◼) starburst sample of 76 are included with more normal spirals. Integrated values for the Galaxy are from 11 and 21 for the FIR. A galactic disk radius of $22.5\mathrm{kpc}$ is assumed. Adapted from 157.

Figure 6

The x-ray halo NGC 5775 from integrated EPIC XMM-Newton observations exhibits substructure correlating on galactic scales with filaments of diffuse ionized gas and spurs in the radiocontinuum. Adapted from 157.

Figure 7

VLA radiocontinuum map of NGC 5775 at $6\mathrm{cm}$. The polarized intensity is given in contours on a gray scale representation of the $\mathrm{H}\alpha$ emission. The bars represent the magnetic-field direction and strength. Adapted from 156, courtesy M. Soida.

Figure 8

Deprojected profiles of the (solid) total $\mathrm{H}\alpha$ surface brightness into (dash-dotted) H II region and (dashed) DIG emission for NGC 7793. From 40.

Figure 9

Steps in the quantification of the total $\mathrm{H}\alpha$ luminosity from the DIG for the representative disk galaxy NGC 157. (a) Continuum subtracted $\mathrm{H}\alpha$ image. (b) Schematic form of H II region catalog, giving position and key to the $\mathrm{H}\alpha$ luminosity of each region. (c) Diffuse $\mathrm{H}\alpha$ map after subtracting off the catalogued H II regions. The brightest H II regions are indicated by circles. (d) Measurement of upper limit for DIG. H II regions are blanked off, then each is assigned a local value of DIG surface brightness. Ellipse shows limit of integrated DIG flux measured. From 171.

Figure 10

Technique used for separating H II region emission from DIG emission. (a) Portion of $\mathrm{H}\alpha$ image of NGC 157 with crowded field (on a spiral arm). Gray scale in emission measure. Circles show mean catalogued radii of H II regions. Dashed line at $73\mathrm{pc}{\mathrm{cm}}^{-6}$, cutoff applied to avoid contamination of DIG by H II regions. (b) Diffuse emission after subtracting off H II regions and applying surface brightness cutoff. (c) Same portion of image in units of $\mathrm{H}\alpha$ surface brightness gradient. Applying uniform cutoff at $12.4\mathrm{pc}{\mathrm{cm}}^{-3}{\text{pixel}}^{-1}$ $({0.28}^{″}∕\mathrm{pix})$ yields H II region boundaries equal to those found in (a), which confirms this separation technique. (d) Map as in (b), but with H II region mask filled at level of local DIG, giving upper limit case for total galaxy DIG luminosity (see text). From 171.

Figure 11

Radial variation of the ratio of integrated DIG luminosity in $\mathrm{H}\alpha$ to the total luminosity for six spiral galaxies (within $0.15R∕R_{25}$ crowding precludes accurate estimates). A canonical value for the integrated ratio of $⩾50\%$ is found, with a total fractional area subtended by the DIG of $\sim 80\%$. From 171.

Figure 12

Comparison of a DIG model with observations. (Left) Modeled surface brightness in $\mathrm{H}\alpha$ of the DIG in the disk of NGC 157 assuming 30% of Lyc photons escape from each H II region, and a simple propagation law through a (macroscopically) uniform slab model for the disk. The result is a projection in the plane of the predicted 3D $\mathrm{H}\alpha$ column density. (Right) Deprojected image of the galaxy with H II regions masked, and a cutoff of $73\times \cos i\mathrm{pc}{\mathrm{cm}}^{-6}$ applied to limit the H II region contamination of the DIG. This shows the basic similarity between these models and the observed DIG, though further refinements are important (see Fig. 13). From 170.

Figure 13

Ratios of observed to modeled DIG of NGC 157, for three variant models of photon escape from H II regions, assuming a slab structure to the H I in the disk. (a) Constant fraction of photons emitted by the O stars escape from each H II region. (b) Only H II regions with luminosities higher than a critical value show significant Lyc photon escape. (c) A constant underlying escape fraction for all H II regions plus an increasing increment for H II regions above a set $\mathrm{H}\alpha$ luminosity. The results are similar for all three. (d) Observed H I column density map (with catalogued H II regions overlaid in black). Note complete coincidence of low zones of observed/modeled ratio with zones of low H I column density, as predicted if photon escape from the H II regions is the main ionizer of the DIG. From 170.

Figure 14

Selected emission line ratios from long slit spectra across the luminous H II region NGC 604 in M33. The solid line is the $\mathrm{H}\alpha$ surface brightness showing rising to the center of the H II region, and falling to low values in the DIG outside it. Note the tendency of the line ratios to rise in the DIG. From 67.

Figure 15

Components of unattenuated and hardened ionizing spectra emerging from a clumpy model H II region, with geometrical filling factor of denser clumps ${10}^{-2}$, clump density $100{\mathrm{cm}}^{-3}$, and mean clump radius $1\mathrm{pc}$. Solid line: Input spectrum emerging diluted but unchanged in relative intensities through interclump medium. Dashed line: Hardened spectrum after transiting internal clumps. In this model, the DIG would be ionized by a combination of both spectra, weighted according to the details of the clump model. For details of these models, see 51.

Figure 16

(Color online) Slices through the ionization structure for point sources in (left) smooth and (right) hierarchically clumped density distributions. The mean vertical density structure for each simulation is that of a Dickey-Lockman disk. From the inner to outer contours, the source luminosities (photons per second) are ${10}^{49}$, $3\times {10}^{49}$, $5\times {10}^{49}$, and ${10}^{50}$. Comparing the contours in the smooth and clumpy models, it is clear that the 3D density structure provides low-density paths allowing photons to reach much larger distances from the ionizing source.

Figure 17

$\mathrm{H}\alpha$ and $\left[\mathrm{N}\mathrm{II}\right]∕\mathrm{H}\alpha$ maps for (upper panels) uniform density and (lower panels) hierarchically clumped H II region models. In each simulation, a uniform background has been added to the $\mathrm{H}\alpha$ and [N II] maps to simulate the H II region being surrounded by low-density DIG. Notice the very irregular boundary in the clumpy model and that the rise of $\left[\mathrm{N}\mathrm{II}\right]∕\mathrm{H}\alpha$ with distance from the central source is much steeper in the smooth model.

Figure 18

Temperature in a shear layer between hot and warm gas. The hot gas (top, $T={10}^6\mathrm{K}$) flows to the left at $10\mathrm{km}{\mathrm{s}}^{-1}$ while the warm gas (bottom, $T=8000\mathrm{K}$) flows to the right at $10\mathrm{km}{\mathrm{s}}^{-1}$. Cooling is not included in this 2D calculation.

Figure 19

Comparison of ionizing flux generated in an evaporating cloud with that generated in a turbulent mixing layer.