We present a study of the [C ii] 157.74 lm fine-structure line in a sample of 15 ultraluminous infrared (IR) galaxies (IR luminosity L IR k10 12 L ; ULIRGs) using the Long Wavelength Spectrometer (LWS) on the Infrared Space Observatory (ISO). We confirm the observed order of magnitude deficit (compared to normal and starburst galaxies) in the strength of the [C ii] line relative to the far-infrared (FIR) dust continuum emission found in our initial report, but here with a sample that is twice as large. This result suggests that the deficit is a general phenomenon affecting 4 out of 5 ULIRGs. We present an analysis using observations of generally acknowledged photodissociation region (PDR) tracers ([C ii], [O i] 63 and 145 lm, and FIR continuum emission), which suggests that a high ultraviolet flux G 0 incident on a moderate density n PDR could explain the deficit. However, comparisons with other ULIRG observations, including CO (1-0), [C i] (1-0), and 6.2 lm polycyclic aromatic hydrocarbon (PAH) emission, suggest that high G 0 =n PDRs alone cannot produce a self-consistent solution that is compatible with all of the observations. We propose that non-PDR contributions to the FIR continuum can explain the apparent [C ii] deficiency. Here, unusually high G 0 and/ or n physical conditions in ULIRGs as compared to those in normal and starburst galaxies are not required to explain the [C ii] deficit. Dust-bounded photoionization regions, which generate much of the FIR emission but do not contribute significant [C ii] emission, offer one possible physical origin for this additional non-PDR component. Such environments may also contribute to the observed suppression of FIR fine-structure emission from ionized gas and PAHs, as well as the warmer FIR colors found in ULIRGs. The implications for observations at higher redshifts are also revisited.

We present the results of Spitzer Infrared Spectrograph low-resolution infrared 5-35 m spectroscopy of nearby ultraluminous infrared galaxies (ULIRGs) at z < 0:15. We focus on the search for the signatures of buried active galactic nuclei (AGNs) in the complete sample of ULIRGs classified optically as non-Seyferts (LINERs or H ii regions). In addition to polycyclic aromatic hydrocarbon (PAH) emission features at 6.2, 7.7, and 11.3 m, the conventional tool of starburst-AGN separation, we use the optical depths of the 9.7 and 18 m silicate dust absorption features to infer the geometry of energy sources and dust at the nuclei of these ULIRGs, namely, whether the energy sources are spatially well mixed with dust (a normal starburst) or are more centrally concentrated than the dust (a buried AGN). Infrared spectra of at least 30%, and possibly 50%, of the observed optical non-Seyfert ULIRGs are naturally explained by emission consisting of (1) energetically insignificant, modestly obscured (A V < 20 30 mag) PAH-emitting normal starbursts and (2) energetically dominant, highly dust-obscured, centrally concentrated energy sources with no PAH emission. We interpret the latter component as a buried AGN. The fraction of ULIRGs showing some buried AGN signatures is higher in LINER ULIRGs than in H ii region ULIRGs. Most of the luminous buried AGN candidates are found in ULIRGs with cool far-infrared colors. Where the absorption-corrected intrinsic AGN luminosities are derivable with little uncertainty, they are found to be of the order of 10 12 L , accounting for the bulk of the ULIRGs' luminosities. The 5-35 m spectroscopic starburst/AGN classifications are generally consistent with our previous classifications based on 3-4 m spectra for the same sample.

The energy sources of nine infrared luminous galaxies (IRLGs) are diagnosed based on their ground-based 3-4 µm spectra. Both the equivalent width of the 3.3 µm polycyclic aromatic hydrocarbon (PAH) emission feature and the 3.3 µm PAH to far-infrared luminosity ratio (L 3.3 /L FIR ) are analyzed. Assuming nuclear compact starburst activity in these sources produces the 3.3 µm PAH emission as strongly as that in starburst galaxies with lower far-infrared luminosities, the followings results are found: For six IRLGs, both the observed equivalent widths and the L 3.3 /L FIR ratios are too small to explain the bulk of their far-infrared luminosities by compact starburst activity, indicating that active galactic nucleus (AGN) activity is a dominant energy source. For the other three IRLGs, while the 3.3 µm PAH equivalent widths are within the range of starburst galaxies, the L 3.3 /L FIR ratios after correction for screen dust extinction are a factor of ∼3 smaller. The uncertainty in the dust extinction correction factor and in the scatter of the intrinsic L 3.3 /L FIR ratios for starburst galaxies do not allow a determination of the ultimate energy sources for these three IRLGs.