We report contributions to cosmic infrared background (CIB) intensities originating from known galaxies and their faint companions at submillimeter wavelengths. Using the publicly-available UltraVISTA catalog, and maps at 250, 350, and 500 {mu}m from
the emph{Herschel} Multi-tiered Extragalactic Survey (HerMES), we perform a novel measurement that exploits the fact that uncatalogued sources may bias stacked flux densities --- particularly if the resolution of the image is poor --- and intentionally smooth the images before stacking and summing intensities. By smoothing the maps we are capturing the contribution of faint (undetected in K_S ~ 23.4) sources that are physically associated, or correlated, with the detected sources. We find that the cumulative CIB increases with increased smoothing, reaching 9.82 +- 0.78, 5.77 +- 0.43, and 2.32 +- 0.19$, rm nW m^{-2} sr^{-1}$ at 250, 350, and 500 {mu}m at 300 arcsec FWHM. This corresponds to a fraction of the fiducial CIB of 0.94 +- 0.23, 1.07 +- 0.31, and 0.97 +- 0.26 at 250, 350, and 500 {mu}m, where the uncertainties are dominated by those of the absolute CIB. We then propose, with a simple model combining parametric descriptions for stacked flux densities and stellar mass functions, that emission from galaxies with log(M/Msun) > 8.5 can account for the most of the measured total intensities, and argue against contributions from extended, diffuse emission. Finally, we discuss prospects for future survey instruments to improve the estimates of the absolute CIB levels, and observe any potentially remaining emission at z > 4.
We quantify the fraction of the cosmic infrared background (CIB) that originates from galaxies identified in the UV/optical/near-infrared by stacking 81,250 (~35.7 arcmin^2) K-selected sources (K_AB < 24.0), split according to their rest-frame U - V
vs. V - J colors into 72,216 star-forming and 9,034 quiescent galaxies, on maps from Spitzer/MIPS (24um), Herschel/SPIRE (250, 350, 500um), Herschel/PACS (100, 160um), and AzTEC (1100um). The fraction of the CIB resolved by our catalog is (69 $pm$ 15)% at 24um, (78 $pm$ 17)% at 70um, (58 $pm$ 13)% at 100um, (78 $pm$ 18)% at 160um, (80 $pm$ 17)% at 250um, (69 $pm$ 14)% at 350um, (65 $pm$ 12)% at 500um, and (45 $pm$ 8)% at 1100um. Of that total, about 95% originates from star-forming galaxies, while the remaining 5% is from apparently quiescent galaxies. The CIB at $lambda$ < 200um is sourced predominantly from galaxies at z < 1, while at $lambda$ > 200um the bulk originates from 1 < z < 2. Galaxies with stellar masses log(M/ M_sun)=9.5-11 are responsible for the majority of the CIB, with those in the log(M/ M_sun)=9.5-10 contributing mostly at $lambda$ < 250um, and those in the log(M/ M_sun)=10.5-11 bin dominating at $lambda$ > 350um. The contribution from galaxies in the log(M/ M_sun)=9.0-9.5 and log(M/ M_sun)=11.0-12.0 stellar mass bins contribute the least, both of order 5%, although the highest stellar-mass bin is a significant contributor to the luminosity density at z > 2. The luminosities of the galaxies responsible for the CIB shifts from a combination of normal and luminous infrared galaxies (LIRGs) at $lambda$ < 160um, to LIRGs at 160um < $lambda$ < 500um, to finally LIRGs and ultra-luminous infrared galaxies (ULIRGs) at $lambda$ > 500um. Stacking analyses were performed with SIMSTACK (available at http://www.astro.caltech.edu/~viero/viero_homepage/toolbox.html) which accounts for possible biases due to clustering.
