Results are presented for [CII] 158 µm line fluxes observed with the Herschel PACS instrument in 112 sources with both starburst and AGN classifications, of which 102 sources have confident detections. Results are compared with midinfrared spectra from the Spitzer Infrared Spectrometer and with L ir from IRAS fluxes; AGN/starburst classifications are determined from equivalent width of the 6.2 µm PAH feature. It is found that the [CII] line flux correlates closely with the flux of the 11.3 µm PAH feature independent of AGN/starburst classification, log [f([CII] 158 µm)/f(11.3 µm PAH)] = -0.22 ± 0.25. It is concluded that [CII] line flux measures the photodissociation region associated with starbursts in the same fashion as the PAH feature. A calibration of star formation rate for the starburst component in any source having [CII] is derived comparing [CII] luminosity L([CII]) to L ir with the result that log SFR = log L([CII)]) -7.08 ± 0.3, for SFR in M ⊙ yr −1 and L([CII]) in L ⊙ . The decreasing ratio of L([CII]) to L ir in more luminous sources (the "[CII] deficit") is shown to be a consequence of the dominant contribution to L ir arising from a luminous AGN component because the sources with largest L ir and smallest L([CII])/L ir are AGN.

A summary is presented for 130 galaxies observed with the Herschel PACS instrument to measure fluxes for the [CII] 158 µm emission line. Sources cover a wide range of active galactic nucleus to starburst classifications, as derived from polycyclic aromatic hydrocarbon (PAH) strength measured with the Spitzer Infrared Spectrograph. Redshifts from [CII] and line to continuum strengths (equivalent width of [CII]) are given for the full sample, which includes 18 new [CII] flux measures. Calibration of L([CII)]) as a star formation rate (SFR) indicator is determined by comparing [CII] luminosities with mid-infrared [NeII] and [NeIII] emission line luminosities; this gives the same result as determining SFR using bolometric luminosities of reradiating dust from starbursts: log SFR = log L([CII)]) -7.0, for SFR in M ⊙ yr −1 and L([CII]) in L ⊙ . We conclude that L([CII]) can be used to measure SFR in any source to a precision of ∼ 50%, even if total source luminosities are dominated by an AGN component. The line to continuum ratio at 158 µm, EW([CII]), is not significantly greater for starbursts (median EW([CII]) = 1.0 µm) compared to composites and AGN (median EW([CII]) = 0.7 µm), showing that the far infrared continuum at 158 µm scales with [CII] regardless of classification. This indicates that the continuum at 158 µm also arises primarily from the starburst component within any source, giving log SFR = log νL ν (158 µm) -42.8 for SFR in M ⊙ yr −1 and νL ν (158 µm) in erg s −1 .

We present spectroscopic results for all galaxies observed with the Spitzer Infrared Spectrograph (IRS) which also have total infrared fluxes f IR measured with the Infrared Astronomical Satellite (IRAS), also using AKARI photometry when available. Infrared luminosities and spectral energy distributions (SEDs) from 8 µm to 160 µm are compared to polycyclic aromatic hydrocarbon (PAH) emission from starburst galaxies or mid-infrared dust continuum from AGN at rest frame wavelengths ∼ 8 µm. A total of 301 spectra are analyzed for which IRS and IRAS include the same unresolved source, as measured by the ratio f ν (IRAS 25 µm)/f ν (IRS 25 µm). Sources have 0.004 < z < 0.34 and 42.5 < log L IR < 46.8 (erg s −1 ) and cover the full range of starburst galaxy and AGN classifications. Individual spectra are provided electronically, but averages and dispersions are presented. We find that log [L IR /νL ν (7.7 µm)] = 0.74 ± 0.18 in starbursts, that log [L IR /νL ν (7.7 µm)] = 0.96 ± 0.26 in composite sources (starburst plus AGN), that log [L IR /νL ν (7.9 µm)] = 0.80 ± 0.25 in AGN with silicate absorption, and log [L IR /νL ν (7.9 µm)] = 0.51 ± 0.21 in AGN with silicate emission. L IR for the most luminous absorption and emission AGN are similar and 2.5 times larger than for the most luminous starbursts. AGN have systematically flatter SEDs than starbursts or composites, but their dispersion in SEDs overlaps starbursts. Sources with the strongest far-infrared luminosity from cool dust components are composite sources, indicating that these sources may contain the most obscured starbursts.

We present a comparison of star formation rates (SFR) determined from mid-infrared 7.7 µm PAH luminosity [SFR(PAH)], from 1.4 GHz radio luminosity [SFR(radio)], and from far ultraviolet luminosity [SFR(UV)] for a sample of 287 starburst galaxies with z < 0.5 having Spitzer IRS observations. The previously adopted relation log [SFR(PAH)] = log [νL ν (7.7 µm)] -42.57±0.2, for SFR in M ⊙ yr −1 and νL ν (7.7 µm) the luminosity at the peak of the 7.7 µm PAH feature in ergs s −1 , is found to agree with SFR(radio). Comparing with SFR(UV) determined independently from ultraviolet observations of the same sources with the GALEX mission (not corrected for dust extinction), the median log [SFR(PAH)/SFR(UV)] = 1.67, indicating that only 2% of the ultraviolet continuum typically escapes extinction by dust within a starburst. This ratio SFR(PAH)/SFR(UV) depends on infrared luminosity, with form log [SFR(PAH)/SFR(UV)] = (0.53±0.05)log [νL ν (7.7 µm)] -21.5±0.18, indicating that more luminous starbursts are also dustier. Using our adopted relation between νL ν (7.7 µm) and L ir , this becomes log [SFR(PAH)/SFR(UV)]= (0.53±0.05)log L ir -4.11±0.18, for L ir in L ⊙ . Only Blue Compact Dwarf galaxies show comparable or greater SFR(UV) compared to SFR(PAH). We also find that the ratio SFR(PAH)/SFR(UV) is similar to that in infrared-selected starbursts for a sample of Markarian starburst galaxies originally selected using optical classification, which implies that there is no significant selection effect in SFR(PAH)/SFR(UV) using starburst galaxies discovered by Spitzer. These results indicate that SFRs determined with ultraviolet luminosities require dust corrections by a factor of ∼ 10 for typical local starbursts but this factor increases to > 700 for the most luminous starbursts at z ∼ 2.5. Application of this factor explains why the most luminous starbursts discovered by Spitzer at z ∼ 2.5 are optically faint; with this amount of extinction, the optical magnitude of a starburst having f ν (7.7 µm) of 1 mJy should be V ∼ 25.6.