Cosmic Infrared Background (CIB)

Updated 29 January 2026

Cosmic Infrared Background (CIB) is the integrated far-infrared to millimeter emission from cosmic dust heated by star formation and AGN, revealing the history of dust-obscured star formation.
Measurement techniques such as absolute photometry, stacking, and fluctuation analyses effectively isolate the CIB signal from dominant foregrounds like zodiacal light and Galactic cirrus.
CIB anisotropies and redshift tomography provide insights into galaxy clustering, dark matter halo properties, and aid cosmological studies including CMB lensing delensing.

The cosmic infrared background (CIB) is the aggregate far-infrared to millimeter-wavelength radiation produced by dust emission in galaxies, integrating the energy released by star formation and active galactic nuclei (AGN) over all cosmic epochs. As observed today, the CIB contains approximately half of the energy radiated by all galaxies since the formation of the first stars. Dust grains in the interstellar media of galaxies absorb ultraviolet and optical stellar light, reradiating the energy thermally at temperatures of T ≈ 10–50 K, which redshifts to a broad peak in the 100–200 μm range. The CIB is a unique probe of the history and distribution of star formation, dust content, and large-scale structure in the Universe. Measurements rely on absolute photometry, source counts, fluctuation analyses, and cross-correlation with large-scale structure tracers, all requiring careful foreground mitigation, particularly with respect to zodiacal light and Galactic cirrus. The CIB's absolute intensity, spectral energy distribution (SED), angular power spectrum, polarization properties, and redshift build-up have been characterized using FIRAS, DIRBE, Herschel, Spitzer, Planck, AKARI, and sounding-rocket experiments such as CIBER, as well as via cross-correlation with galaxy catalogs.

1. Physical Origin and Cosmological Significance

The CIB traces the total infrared emission produced by cosmic processes, predominantly dust-obscured star formation and, at some wavelengths, obscured AGN activity. As the baryonic content of galaxies cycles through stars and ISM, ultraviolet and optical photons are partly absorbed by interstellar dust and re-emitted in the far-infrared. The observed CIB intensity at frequency ν is given by the line-of-sight integral

$I_\nu = \int_0^\infty dz\, \frac{d\chi}{dz}\, a(z)\, j(\nu, z)$

where $j(\nu, z)$ is the comoving volume emissivity (energy per unit time, frequency, comoving volume, and solid angle) and $a=1/(1+z)$ . Approximately half the total extragalactic background light (EBL) energy is in the CIB, reflecting the fact that, globally, about half of all cosmic star formation is dust-obscured.

Fluctuations and anisotropies in the CIB encode the spatial distribution and clustering of dusty galaxies, adding a unique tomographic probe of large-scale structure and the star formation history inaccessible to direct source detection at the faintest fluxes (Berta et al., 2011, Collaboration et al., 2013).

2. Measurement Techniques and Foreground Removal

Direct absolute photometry of the CIB is challenging due to dominant foregrounds: zodiacal light from interplanetary dust, Galactic cirrus from Milky Way dust, and the cosmic microwave background (CMB) at the longest wavelengths. The CIB has been mapped and studied using a combination of the following approaches:

Absolute Photometry: COBE/FIRAS and DIRBE provided early, full-sky absolute SEDs between ≈ 60 μm and 1000 μm (Béthermin et al., 2011, Fixsen et al., 2011). The FIRAS spectrum peaks at λ ≈ 150 μm with $I_\nu(150\,\mu{\rm m}) \approx 15$ –20 nW m $^{-2}$ sr $^{-1}$ .
Number Counts and Stacking: Integration of the resolved source counts $dN/dS$ down to the confusion or detection limit provides robust CIB lower limits at λ ≲ 100 μm. Stacking (e.g., matching 24 μm Spitzer sources to far-IR maps) bridges the gap between resolved counts and the confusion limit, enabling the recovery of up to ≈90% of the CIB at peak wavelengths (Berta et al., 2010, Berta et al., 2011).
Fluctuation Analyses ( $P(D)$ ): Probability of deflection methods model the pixel flux distribution to constrain counts below the confusion threshold, statistically recovering the contribution of faint sources (Berta et al., 2011).
Foreground Mitigation: Zodiacal light is ∼10³ times brighter than the CIB below 30 μm; uncertainties in Zodiacal subtraction dominate systematic errors in the mid-IR. At λ ≳ 60 μm, Galactic cirrus becomes the main contaminant and is removed via high-resolution H I templates or fitting its power spectrum (Pénin et al., 2011, Lenz et al., 2019). For CMB foregrounds at λ ≳ 500 μm, subtraction is straightforward due to the well-known CMB SED.
Cross-Correlation and Tomography: Cross-correlating CIB intensity maps with spectroscopic galaxy catalogs or quasars enables tomographic recovery of the redshift distribution of the CIB without requiring individual source detection (Schmidt et al., 2014, Chiang et al., 7 Apr 2025, Yan et al., 2023).

3. Spectral Energy Distribution and Intensity Budget

The absolute SED of the CIB from ≈ 8–1000 μm, when foregrounds are mitigated, is well-constrained to ≈20–30% in most of the range (Béthermin et al., 2011, Chiang et al., 7 Apr 2025). Key measurements include:

Wavelength (μm)	I_CIB (nW m⁻² sr⁻¹)	Instrument(s)
65	12.5 ± 1.4 ± 9.2	AKARI
90	22.3 ± 1.7 ± 4.7	AKARI
100	14.4 ± 6.0	DIRBE/FIRAS, Herschel
140	12.4 ± 6.9	DIRBE/FIRAS
160	14.4 ± 3.0	Spitzer/MIPS
250	10.5 ± 1.6	Herschel/SPIRE
350	6.7 ± 1.5	Herschel/SPIRE
500	3.1 ± 0.7	Herschel/SPIRE

Recovered intensities from stacking and $P(D)$ analyses reach ≈90% of the background at the peak. Resolved-source integration alone accounts for ≈45–75% at 100–160 μm, with stacking and $P(D)$ raising this to ≳90% (Berta et al., 2010, Berta et al., 2011, Duivenvoorden et al., 2019). At long wavelengths (250–500 μm), map-fitting techniques with deep multiwavelength priors indicate that with current catalogs, the CIB can be resolved to ≳90% at λ ≤ 500 μm (Duivenvoorden et al., 2019).

The CIB SED is well-modeled by a modified blackbody $I_\nu \propto \nu^{3+\beta}/[\exp(h\nu/kT_d)-1]$ with a dust temperature evolving from T_d ≈ 20 K at z = 0 to ≈ 30–40 K at z ≈ 2–4, and dust emissivity index β ≈ 1.5–2.0 (Collaboration et al., 2013, Yan et al., 2023).

4. Anisotropies, Clustering, and Polarization

Spatial fluctuations in the CIB encode cosmological information on the clustering of star-forming galaxies and their host dark-matter halos. The angular power spectrum $C_\ell$ combines shot noise (Poisson) and clustering (2-halo and 1-halo) components. Planck, Herschel, and Spitzer have measured the CIB $C_\ell$ from ℓ ≈ 70 to ℓ ≈ 3000, with power laws of the form $C_\ell \propto \ell^{-1.2}$ (Lenz et al., 2019, Collaboration et al., 2013).

The bispectrum Bℓ1ℓ2ℓ3, sensitive to the non-Gaussianity of the dusty galaxy distribution, shows a steep power law with increasing contribution from massive halos (Collaboration et al., 2013).

The polarization of CIB anisotropies, predicted from the intrinsic alignment (IA) of galactic dust emission, is suppressed by ≈10⁻⁴ relative to the intensity power and is observationally negligible at CMB frequencies (≤353 GHz), but may become detectable at submm frequencies with next-generation missions (Feng et al., 2019).

5. Redshift Distributions and Dust-Obscured Star Formation

The build-up of the CIB as a function of redshift—critical for mapping cosmic star-formation history—has been reconstructed using photometric/spectroscopic redshifts in deep multiwavelength fields, stacking, and cross-correlation tomography. Key results include:

At 70 and 100 μm, ≳80% of the CIB is produced at z ≤ 1.0; at 160 μm, this drops to ≈55%, with the half-light redshift shifting from z ≈ 0.6 at 70 μm to z ≈ 0.73 at 160 μm (Berta et al., 2011, Jauzac et al., 2010).
The CIB intensity as a function of z peaks at z ≈ 1–1.2, accurately tracing the dust-obscured peak of cosmic star formation (Schmidt et al., 2014).
Tomographic intensity mapping with Planck, Herschel, and IRAS, cross-correlated with SDSS/BOSS/eBOSS, recovers the evolving CIB spectrum and total infrared luminosity density to 0.04 dex precision over 0 < z < 4 (Chiang et al., 7 Apr 2025). The cosmic dust density Ω_dust(z) peaks at z ≈ 1–1.5 and declines by a factor ≈3 to the present.
Dust-obscured star formation dominates: at z ≈ 2, ≈90% of the total SFR density is in the IR-inferred, dust-obscured component, with ≈80% at z = 0 and ≈60% even at z = 4 (Chiang et al., 7 Apr 2025).

6. Connection to Large-Scale Structure and Cosmology

CIB fluctuations are a tracer of the high-redshift large-scale structure and cosmic matter distribution. Because most CIB emission traces star-forming galaxies residing in dark matter halos of 10¹¹–10¹³ M⊙ at z ≈ 1–2, the CIB is significantly correlated with tracers such as CMB gravitational lensing convergence κ. The cross-correlation CIB–κ is exploited for sample-variance cancellation and improvement of astrophysical model constraints (McCarthy et al., 2020, Lenz et al., 2019). The CIB is now used to "delens" the CMB and to improve primordial B-mode searches.

CIB–galaxy cross-correlations simultaneously constrain the star-formation rate, dust SED, and halo occupation distribution (HOD). The peak efficiency of dusty star formation occurs in halos of mass ≈10¹² M⊙, with dust temperature T₀ ≈ 21–24 K at z = 0, rising moderately with redshift (Yan et al., 2023). These analyses yield galaxy linear bias evolution in agreement with independent clustering results.

Angular power and bispectrum analyses allow tests of primordial non-Gaussianity and cosmological parameters. Large-scale CIB fluctuations are sensitive to primordial scale-dependent bias (Lenz et al., 2019, Collaboration et al., 2013).

7. Source-Subtracted Anisotropies and Faint Population Constraints

After subtraction of resolved sources, large-scale CIB fluctuations in the near-IR (1–5 μm) exceed the level expected from remaining normal galaxies by factors of several, indicating contributions from constituent populations below current survey depth. These fluctuations are highly coherent with the unresolved cosmic X-ray background, consistent with a significant black hole accretion component (Kashlinsky et al., 2019). The spectral shape and coherence provide evidence for new high-z populations, potentially including Population III stars, direct-collapse black holes, and other exotic sources (Kashlinsky et al., 2018).

Upcoming wide and deep surveys (Euclid, LSST, WFIRST, JWST, Athena, eROSITA) will expand CIB fluctuation and tomographic studies, enabling measurements of the earliest star formation, black hole growth, and baryonic feedback processes at redshifts inaccessible to direct observation.

References: