Dust spectroscopy provides the most diagnostic information on the dust composition. Our knowledge about the composition of the dust in the Galactic diffuse ISM is mainly derived from the absorption and emission spectral lines: the 2175 Å extinction bump (small graphitic dust), the 3.4 µm absorption feature (aliphatic hydrocarbon dust), the 9.7 µm and 18 µm absorption features (amorphous silicate dust), and the 3.3, 6.2, 7.7, 8.6, and 11.3 µm emission features (polycyclic aromatic hydrocarbon [PAH] molecules). The ice absorption features at 3.1 and 6.0 µm (H2O), 4.67 µm (CO), 4.27 and 15.2 µm (CO2), 3.54 and 9.75 µm (CH3OH), 2.97 µm (NH3), 7.68 µm (CH4), 5.81 µm (H2CO), and 4.62 µm (XCN-) are seen in dark molecular clouds with visual extinction AV > 3 mag. In this section I will present a comparative overview of the dust absorption and emission features in AGNs and the inferred dust composition.
3.1. The 2175 Å Extinction Bump
The 2175 Å extinction bump, first detected over 40 years ago (Stecher 1965), is an ubiquitous feature of the Milky Way ISM. With a stable central wavelength and variable feature strength for lines of sight in our Galaxy, the 2175 Å bump is relatively weaker in the LMC and absent in the SMC (see Fig. 2). This bump is largely absent in AGNs (see Section 2) except Gaskell & Benker (2007) recently claimed that it might be detected in Mrk 304, one of the seven AGNs with the highest quality extinction curves in their 14-AGN sample. Fig. 4 compares the UV spectra of 5 slightly reddened type 1 AGNs with the template of type 1 AGNs reddened with the standard Galactic extinction. It is seen that the Galactic extinction predicts too strong a 2175 Å dip (Maiolino et al. 2001b).
Figure 4. Comparison of the UV spectra (solid lines) of 5 type 1 AGNs (whose broad line ratios and continuum suggest dust absorption) with the template of type 1 AGNs reddened by a standard Galactic extinction (RV = 3.1) with a total amount of reddening E(B - V) consistent with that inferred from the broad lines and adapted to match the shape of the continuum in those regions not affected by the 2175 Å bump (dashed line). Taken from Maiolino et al. (2001a) with modifications.
The exact nature of the 2175 Å bump, the strongest spectroscopic extinction feature in the Galactic ISM, remains uncertain. It is generally believed to be caused by aromatic carbonaceous (graphitic) materials, very likely a cosmic mixture of PAH molecules (Joblin et al. 1992; Li & Draine 2001b). The fact that the 2175 Å bump is not (or at least rarely) seen in AGNs suggests that its carrier (e.g. PAHs) may have been photodestroyed by energetic photons (e.g. X-ray irradiation) from the central engine.
3.2. The 3.4 µm Aliphatic Hydrocarbon Absorption Feature
The 3.4 µm absorption feature, attributed to the C-H stretching mode in saturated aliphatic hydrocarbon dust, is widely seen in the Galactic diffuse ISM (but never seen in molecular clouds; see Pendleton & Allamandola 2002). This feature is also seen in AGNs (Wright et al. 1996, Imanishi et al. 1997, Mason et al. 2004), closely resembling that of our Galaxy in both peak wavelengths and relative feature strengths of the 3.42 µm, 3.48 µm, and 3.51 µm subfeatures (corresponding to symmetric and asymmetric stretches of C-H bonds in CH2 and CH3 groups in aliphatic hydrocarbon chains). Mason et al. (2004) argued that the 3.4 µm absorption feature at least in face-on Seyfert 2 galaxies arises in dust local to the active nucleus rather than in the diffuse ISM of the galaxy.
The exact carrier of this feature remains uncertain. So far, among the > 20 candidate materials proposed over the years since its first detection in the Galactic center sightlines 28 years ago, the experimental spectra of hydrogenated amorphous carbon (Mennella et al. 1999) and the organic refractory residue, synthesized from UV photoprocessing of interstellar ice mixtures (Greenberg et al. 1995), provide the best fit to the observed spectra.
So far, no polarization has been detected for this feature (Adamson et al. 1999, Chiar et al. 2006, Mason et al. 2006), suggesting that the carrier of this feature is either spherical or unaligned or both. Spectropolarimetric measurements for both the 9.7 µm silicate and the 3.4 µm hydrocarbon features for the same sightline (e.g. Chiar et al. 2006) would allow for a direct test of the silicate core-hydrocarbon mantle interstellar dust model (Li & Greenberg 1997, Jones et al. 1990), since this model predicts that the 3.4 µm feature would be polarized if the 9.7 µm feature (for the same sightline) is polarized (Li & Greenberg 2002).
3.3. The 9.7 µm and 18 µm Silicate Absorption and Emission Features
The strongest IR absorption features in the Galactic ISM are the 9.7 µm and 18 µm bands, which are almost certainly due to silicate minerals: they are respectively ascribed to the Si-O stretching and O-Si-O bending modes in some form of silicate material (e.g. olivine Mg2xFe2-2xSiO4). The observed interstellar silicate bands are broad and relatively featureless, indicating that interstellar silicates are largely amorphous rather than crystalline (Li & Draine [2001a] estimated that the amount of a < 1 µm crystalline silicate grains in the Galactic diffuse ISM is < 5% of the solar Si abundance).
The first detection of the silicate absorption feature in AGNs was made at 9.7 µm for the prototypical Seyfert 2 galaxy NGC 1068 (Rieke & Low 1975; Kleinmann et al. 1976), indicating the presence of a large column of silicate dust in the line-of-sight to the nucleus. It is known now that most of the type 2 AGNs display silicate absorption bands (e.g. see Roche et al. 1991, Siebenmorgen et al. 2004) as expected - for a centrally heated optically thick torus viewed edge-on, the silicate features should be in absorption. Spatially resolved mid-IR spectra obtained for NGC 1068 (Mason et al. 2006, Rhee & Larkin 2006) and Circinus (Roche et al. 2006) have revealed striking variations in continuum slope, silicate feature profile and depth.
However, it appears that the 9.7 µm silicate absorption profile of AGNs differs from that of the Milky Way. Jaffe et al. (2004) found that the 9.7 µm silicate absorption spectrum of NGC 1068 shows a relatively flat profile from 8 to 9 µm and then a sharp drop between 9 and 10 µm; in comparison, the Galactic silicate absorption profiles begin to drop already at ~ 8 µm. They obtained a much better fit to the 9.7 µm absorption feature of NGC 1068 by using the profile of calcium aluminium silicate Ca2Al2SiO7, a high-temperature dust species found in some supergiant stars (Speck et al. 2000). It would be interesting to know if the amount of calcium required to account for the observed absorption is consistent with abundance constraints. Very recently, Roche et al. (2007) reported the detection of a spectral structure near 11.2 µm in NGC 3094, indicative of the possible presence of crystalline silicates in AGNs.
For type 1 AGNs viewed face-on, one would expect to see the silicate features in emission since the silicate dust in the surface of the inner torus wall will be heated to temperatures of several hundred kelvin by the radiation from the central engine, allowing for a direct detection of the 9.7 µm and 18 µm silicate bands emitted from this hot dust. However, their detection (using Spitzer) has only very recently been reported in a number of type 1 AGNs (Hao et al. 2005, Siebenmorgen et al. 2005, Sturm et al. 2005, Weedman et al. 2005, Shi et al. 2006). Siebenmorgen et al. (2005) postulated that the AGN luminosity determines whether the silicate emission bands are prominent or not (i.e., they may be present only in the most luminous AGNs), but this idea was challenged by their detection in the low-luminosity AGN NGC 3998, a type 1 LINER galaxy (Sturm et al. 2005).
The 9.7 µm silicate emission profiles of both quasars (high luminosity counterparts of Seyfert 1 galaxies; Hao et al. 2005, Siebenmorgen et al. 2005) and the low-luminosity AGN NGC 3998 (Sturm et al. 2005) peak at a much longer wavelength (~ 11 µm), inconsistent with "standard" silicate ISM dust (which peaks at ~ 9.7 µm). The 9.7 µm feature of NGC 3998 is also much broader than that of the Galactic ISM (Sturm et al. 2005). The deviations of the silicate emission profiles of type 1 AGNs from that of the Galactic ISM dust may indicate differences in the dust composition, grain size distribution, or radiative transfer effects (Sturm et al. 2005, Levenson et al. 2007). The red tail of the 18 µm silicate feature of NGC 3998 is significantly weaker than that of the bright quasars (Sturm et al. 2005), suggesting that there may exist significant environmental variations. Finally, it is worth noting that the 9.7 µm silicate feature of Mkn 231, a peculiar type 1 Seyfert galaxy, is also seen in absorption peaking at ~ 10.5 µm (Roche et al. 1983).
3.4. The 3.3, 6.2, 7.7, 8.6 and 11.3 µm PAH Emission Features
The distinctive set of "Unidentified Infrared" (UIR) emission features at 3.3, 6.2, 7.7, 8.6, and 11.3 µm, now generally identified as the vibrational modes of PAH molecules (Léger & Puget 1984; Allamandola et al. 1985), are seen in a wide variety of Galactic and extragalactic regions (see Draine & Li 2007). In the Milky Way diffuse interstellar medium (ISM), PAHs, containing ~ 45 ppm (parts per million, relative to H) C, account for ~ 20% of the total power emitted by interstellar dust (Li & Draine 2001b). The ISO (Infrared Space Observatories) and Spitzer imaging and spectroscopy have revealed that PAHs are also a ubiquitous feature of external galaxies. Recent discoveries include the detection of PAH emission in a wide range of systems: distant Luminous Infrared Galaxies (LIRGs) with redshift z ranging from 0.1 to 1.2 (Elbaz et al. 2005), distant Ultraluminous Infrared Galaxies (ULIRGs) with redshift z ~ 2 (Yan et al. 2005), distant luminous submillimeter galaxies at redshift z ~ 2.8 (Lutz et al. 2005), elliptical galaxies with a hostile environment (containing hot gas of temperature ~ 107 K) where PAHs can be easily destroyed through sputtering by plasma ions (Kaneda et al. 2005), faint tidal dwarf galaxies with metallicity ~ Z / 3 (Higdon et al. 2006), and galaxy halos (Irwin & Madden 2006, Engelbracht et al. 2006).
However, the PAH features are absent in AGNs, as first noticed by Roche et al. (1991). This is commonly interpreted as the destruction of PAHs by extreme UV and soft X-ray photons in AGNs (Roche et al. 1991; Voit 1991, 1992; Siebenmorgen et al. 2004). Genzel et al. (1998) proposed to use the line-to-continuum ratio of the 7.7 µm PAH feature as a discriminator between starburst and AGN activity in ULIRGs (i.e. whether the dominant luminosity source of ULIRGs is an AGN or a starburst). We should note that the PAH emission features are detected in some Seyfert 2 galaxies, but they are from the circumnuclear star-forming regions, not from the AGNs (e.g. see Le Floc'h et al. 2001, Siebenmorgen et al. 2004).
3.5. The Ice Absorption Features
Grains in dark molecular clouds (usually with AV > 3 mag) obtain ice mantles consisting of H2O, NH3, CO, CH3OH, CO2, CH4, H2CO and other molecules (with H2O as the dominant species), as revealed by the detection of various ice absorption features (e.g., H2O: 3.1, 6.0 µm; CO: 4.67 µm; CO2: 4.27, 15.2 µm; CH3OH: 3.54, 9.75 µm; NH3: 2.97 µm; CH4: 7.68 µm; H2CO: 5.81 µm; XCN-: 4.62 µm). The ice absorption features are also seen in most ULIRGs (e.g. see Spoon et al. 2002), indicating the presence of a large quantity of molecular material in ULIRGs. However, the ice absorption features are not expected in AGNs due to the high dust temperatures (because of the immense bolometric luminosity emitted from the AGN) - the dust in the torus, even at a distance of ~ 100 pc, is too warm (> 100 K) for ice mantles to survive.