Infrared-bright galaxies such as those discussed in the previous section are hosts to large quantities of molecular gas and dust (e.g., Solomon et al. 1997). The optical line ratios measured in these objects are undoubtedly affected by dust extinction. It is therefore important to also observe these objects at longer wavelengths to verify the results derived from the optical spectra. Near-infrared spectroscopy has had success finding obscured BLRs in several ULIGs (e.g., Hines 1991; Veilleux et al. 1997b, 1999b; spectropolarimetry has lent support to some of these findings: Hines 1991; Hough et al. 1991; Hines & Wills 1993; Hines et al. 1995; Young et al. 1993). This technique has also proven useful in the study of highly reddened BLRs in intermediate Seyferts (1.8's and 1.9's; Goodrich 1990; Rix et al. 1990) and in optically classified Seyfert 2 and radio galaxies (e.g., Blanco, Ward, & Wright 1990; Goodrich, Veilleux, & Hill 1994; Ruiz, Rieke, & Schmidt 1994; Hill, Goodrich, DePoy 1996; Veilleux, Goodrich, & Hill 1997a).
The line of choice for ground-based near-infrared searches of obscured BLRs in nearby galaxies is Pa at 1.8751 µm (Table 4). Under Case B recombination (Osterbrock 1989), this line is one-third the strength of H and is twelve times stronger than Br 2.1655, the next best diagnostic line (e.g., Goldader et al. 1995). This huge gain in intensity more than compensates the slightly larger optical depth due to extinction at the shorter wavelength of Pa (see 4th column in Table 4).
Another important AGN diagnostic line in the K band is [Si VI] 1.962. The existence of five-times ionized silicon ions requires energies larger than 167 eV (Table 4). This forbidden line has been detected in a number of optically selected Seyfert 2 galaxies with a strength comparable to that of [Fe VII] 6087 (roughly a tenth the strength of H), as expected from photoionization by a AGN power-law continuum (Oliva & Moorwood 1990; Greenhouse et al. 1993; Marconi et al. 1994; Oliva et al. 1994; Thompson 1995, 1996). Near-infrared spectroscopic surveys of ULIGs have confirmed the optical results: the fraction of objects with genuine AGNs (with a BLR or strong [Si VI] 1.962 feature) is at least ~ 20 - 25%, but reaches ~ 35 - 50% for those objects with log[Lir / L] > 12.3. Nevertheless, the presence of an AGN in ULIGs does not necessarily imply that AGN activity is the dominant source of energy in these objects. A more detailed look at the AGNs in these ULIGs is needed to answer this question.
|Hydrogen Recombination Lines||High-Ionization Lines|
|Line||(µm)||F/FH||A / AH||Line||(eV)|
|H||0.4861||1.00||1.48||[S IX] 1.252 µm||328|
|H||0.6563||2.85||1.00||[Si X] 1.430 µm||351|
|Pa||1.0938||0.090||0.45||[Si XI] 1.932 µm||401|
|Pa||1.2818||0.162||0.34||[Si VI] 1.962 µm||167|
|Pa||1.8751||0.332||0.18||[Ca VIII] 2.321 µm||128|
|Br||2.1655||0.0275||0.14||[Si VII] 2.483 µm||205|
|Br||4.0512||0.0779||0.05||[Si IX] 3.935 µm||303|
A strong linear correlation has long been known to exist between the continuum (or, equivalently, bolometric) luminosities of broad-line AGN and their emission-line luminosities (e.g., Yee 1980; Shuder 1981; Osterbrock 1989). This correlation has often been used to argue that the broad-line regions in AGNs are photoionized by the nuclear continuum. If this is the case, the broad-line-to-bolometric luminosity ratio is a measure of the covering factor of the BLR (e.g., Osterbrock 1989). This correlation can be used to estimate the importance of the AGN in powering ultraluminous infrared galaxies (Veilleux et al. 1997b, 1999b). In ULIGs powered uniquely by an AGN, we expect the broad-line luminosities to fall along the correlation for AGNs. Any contribution from a starburst will increase the bolometric luminosity of the ULIG without a corresponding increase in the broad-line luminosity. Starburst-dominated ULIGs are therefore expected to fall below the "pure-AGN" correlation traced by the optical quasars in a diagram of LH(BLR) plotted as a function of Lbol. The data of ULIGs with optical and obscured BLRs are shown in Figure 3. A discussion of the methods and assumptions which were used to create this figure is presented in Veilleux et al. (1999a). Figure 3 strongly suggests that most (~ 80%) of the ULIGs with optical or near-infrared BLRs in the 1-Jy sample are powered predominantly by the quasar rather than by a powerful starburst. In other words, the detection of an optical or near-infrared BLR in a ULIG (about 20% of the total 1-Jy sample) appears to be an excellent sign that the AGN is the dominant energy source in that ULIG.
Figure 3. Dominant energy source of ultraluminous infrared galaxies based on their broad-line luminosities. The solid line is the best fit for the optical quasars. From Veilleux et al. (1999b).