Luminosity functions of local infrared galaxies with AKARI: implications for the cosmic star formation history and AGN evolution

Abstract

Infrared (IR) luminosity is fundamental to understanding the cosmic star formation history
\nand active galactic nuclei (AGN) evolution, since their most intense stages are often obscured
\nby dust. However, local IR luminosity function estimates today are still based on the IRAS
\nsurvey in the 1980s, with wavelength coverage only up to 100 μm. The AKARI IR space
\ntelescope performed an all-sky survey in six IR bands (9, 18, 65, 90, 140 and 160 μm) with
\n3–10 times better sensitivity, covering the crucial far-IR wavelengths across the peak of the
\ndust emission. Combined with a better spatial resolution, AKARI can much more precisely
\nmeasure the total infrared luminosity (L_(TIR)) of individual galaxies, and thus, the total infrared
\nluminosity density in the local Universe.
\nBy fitting modern IR spectral energy distribution (SED) models, we have remeasured L_(TIR)
\nof the IRAS Revised Bright Galaxy Sample, which is a complete sample of local galaxies with
\nS_(60μm) > 5.24 Jy.
\nWe present mid-IR monochromatic luminosity (νL_ν) to L_(TIR) correlations for Spitzer 8 μm,
\nAKARI 9 μm, IRAS 12 μm, WISE 12 μm, ISO 15 μm, AKARI 18 μm, WISE 22 μm and Spitzer
\n24 μm filters. These measures of L_(MIR) are well correlated with L_(TIR), with scatter in the range
\n13–44 per cent. The best-fitting L_(MIR)-to-L_(TIR) conversions provide us with estimates of L_(TIR)
\nusing only a single MIR band, in which several deep all-sky surveys are becoming available
\nsuch as AKARI MIR and WISE.
\nAlthough we have found some overestimates of L_(TIR) by IRAS due to contaminating cirrus/
\nsources, the resulting AKARI IR luminosity function (LF) agrees well with that from
\nIRAS. We integrate the LF weighted by L_(TIR) to obtain a cosmic IR luminosity density of
\nΩ_(TIR) = (8.5^(+1.5)_(−2.3)) × 10^7 L_⊙ Mpc^(−3), of which 7 ± 1 per cent is produced by luminous infrared galaxies (LIRGs) (L_(TIR) > 10^(11) L_⊙), and only 0.4 ± 0.1 per cent is from ultraluminous infrared
\ngalaxies (ULIRGs) (L_(TIR) > 10^(12) L_⊙) in the local Universe, in stark contrast to high-redshift
\nresults.
\nWe separate the contributions from AGN and star-forming galaxies (SFGs). The SFG IR LF
\nshows a steep decline at the bright end. Combined with high-redshift results from the AKARI
\nNEP deep survey, these data show a strong evolution of Ω^(SF)_(TIR)
\n∝ (1 + z)^(4.0 ± 0.5) and Ω^(AGN)_(TIR)
\n∝
\n(1 + z)^(4.4 ± 0.4). For Ω^(AGN)_(TIR), the ULIRG contribution exceeds that from LIRGs already by z ~ 1.
\nA rapid evolution in both Ω^(AGN)_(TIR) and Ω^(SFG)_(TIR) suggests the correlation between star formation and
\nblack hole accretion rate continues up to higher redshifts. We compare the evolution of Ω^(AGN)_(TIR)
\nto that of X-ray luminosity density. The Ω^(AGN)_(TIR)/Ω^(AGN)_(X-ray) ratio shows a possible increase at z >
\n1, suggesting an increase of obscured AGN at z > 1.