The Cosmic Microwave Background’s Detailed Structure

The early universe contained a hot plasma. Photons scattered frequently from free electrons, which were coupled to the nuclei through electromagnetic interactions. Radiation and ordinary matter therefore behaved, to a useful approximation, as a shared photon–baryon fluid. The radiation pressure resisted gravitational compression.[1]

As expansion cooled the plasma, electrons and protons increasingly combined into neutral hydrogen. This process, recombination, reduced the free-electron population and made scattering much less frequent. The main last-scattering period occurred around 380,000 years after the Big Bang, at a temperature near 3,000 K.[1]

The transition had a finite duration. The last-scattering surface is the region from which that ancient light reaches us now, with some thickness in time and distance. It is not a physical wall or the edge of the universe. Expansion subsequently stretched the radiation to microwave wavelengths, and some photons scattered again later.[1]

The CMB’s average spectrum closely follows a blackbody: the characteristic distribution of radiation in thermal equilibrium. A widely used determination gives T = 2.7255 ± 0.0006 K. This precise value comes from later measurement and calibration work, not from the original discovery experiment.[2]

COBE’s FIRAS instrument established the extraordinary closeness of the spectrum to a blackbody. Limits on departures constrain the universe’s thermal history and the amount and timing of energy release. They do not imply that no energy was ever added afterward, or that the radiation experienced no later interactions.[3]

The colors on a CMB map are deliberately exaggerated

After accounting for the large dipole pattern associated mainly with our motion, the primary temperature variations are of order one part in 100,000, with amplitudes of tens of microkelvins depending on angular scale. An anisotropy is simply a difference with direction. A warmer patch can reflect the local radiation temperature, a Doppler contribution from motion, gravitational effects, or their combination.[4]

In 1965, Arno Penzias and Robert Wilson reported unexplained excess antenna noise at 4.08 GHz, corresponding to 3.5 ± 1 K. Their discovery opened the observational story of the CMB; the much more precise modern temperature was established later.[5]

Planck observed in nine frequency bands, with angular resolution ranging roughly from 33 to 5 arcminutes depending on the band. Its temperature, polarization, and lensing products remain major reference datasets. Different observing frequencies supplied a way to distinguish cosmic radiation from foreground emission.[8]

Dust, synchrotron radiation from our Galaxy, distant sources, and instrumental noise all enter the measurements. Researchers combine frequency maps and test independent separation methods to recover the CMB component. A cleaned map is therefore a carefully evaluated data product, with residual uncertainties that must accompany its interpretation.[9]

Before decoupling, gravity compressed regions of the photon–baryon fluid while radiation pressure drove them outward. Different wavelengths underwent different stages of compression and rarefaction before the light last scattered. These coherent acoustic oscillations produced the characteristic sequence of peaks. They were physical waves in an early plasma; the plotted peaks are not separate explosions.[10]

From patches on the sky to an angular spectrum

The multipole ℓ labels angular structure: lower values correspond to broad patterns, higher values to finer detail. A useful approximate scale is θ ≈ 180°/ℓ. The temperature power spectrum C_ℓ^TT describes the statistical strength at each multipole; plots commonly show the rescaled quantity D_ℓ^TT = ℓ(ℓ + 1)C_ℓ^TT/(2π).[4]

Peak heights carry several kinds of information. Added baryons increase the fluid’s inertia and change the balance between compression and rarefaction peaks. Dark matter and radiation affect the gravitational potentials and expansion history. Researchers interpret the whole pattern together, rather than assigning each peak to one independent ingredient.[10]

On smaller scales, photons diffuse through the plasma and smooth temperature differences, producing diffusion damping, often called Silk damping. The finite duration of last scattering also suppresses fine angular structure. A falling high-ℓ spectrum is part of the physical signal, not solely a limitation of telescope resolution.[12]

Linear polarization describes a preferred orientation of the light’s electric-field oscillations. Thomson scattering can generate it when the radiation arriving at an electron has a quadrupole pattern: a particular directional imbalance. Astronomers combine polarization orientations across the sky into E-mode and B-mode patterns; a single measured orientation cannot be labeled E or B on its own.[14]

DASI reported the first detection of CMB polarization, including E modes, in 2002. WMAP’s early achievements included temperature–polarization correlations and a large-angle signal associated with reionization.[15], [16]

Reionization leaves a later polarization signature

When early luminous objects ionized intergalactic gas, newly freed electrons scattered some CMB photons. The resulting large-angle polarization helps constrain the optical depth τ, which measures the integrated scattering along the journey. Different ionization histories can produce similar τ, so it does not provide a unique date for the first stars or an instantaneous reionization event.[17]

Observed B modes and the primordial signal still sought

Scalar perturbations generate E modes but no B modes at linear order before later remapping. Gravitational lensing converts some of the apparent E pattern into B. SPTpol detected lensing B modes in 2013; their existence is established even though a primordial gravitational-wave contribution remains unconfirmed.[18]

Polarized Galactic dust and synchrotron emission also contribute. Observations at multiple frequencies separate those foregrounds statistically, while delensing reduces the contribution caused by gravitational remapping. These address different sources of confusion in the search for primordial B modes.[19]

Flat ΛCDM includes a cosmological constant and cold dark matter. Six cosmological parameters determine its predicted temperature and polarization spectra.[13]

A scalar index slightly below one describes a nearly scale-invariant spectrum with somewhat more relative power at larger scales. Together with other properties of the initial fluctuations, this agrees with many inflationary models. It does not uniquely identify an inflaton, establish its energy scale, or prove that inflation is the only possible early history.[20]

Precision depends on the question and the assumptions

Different combinations of expansion history and curvature can produce similar angular scales in the CMB. Galaxy BAO measurements help separate these possibilities. Curvature constraints therefore belong to a specified model and combination of observations; spatial flatness is an assumption of the six-parameter baseline, not an extra parameter it independently measures.[21]

The CMB also tests the early relativistic energy density and the effects of neutrinos. The parameter N_eff describes radiation content beyond photons, rather than simply counting detected neutrino particles. Neutrino-mass limits depend on the cosmological model and additional observations. ACT’s extended-model analyses illustrate why those assumptions must accompany any quoted bound.[22]

Gravitational lensing slightly deflects the light as it travels through intervening structure. The resulting correlations allow researchers to reconstruct a projected lensing field and test the growth of structure. Modern SPT analyses use this information alongside temperature and polarization; a reconstruction combines material along sightlines rather than resolving every dark matter structure in three dimensions.[23]

The integrated Sachs–Wolfe effect changes photon energies when gravitational potentials evolve during the journey. At late times, this connects large-angle temperature patterns with the evolution of structure and accelerated expansion. Cross-correlations with galaxy maps help distinguish the contribution from the much larger primary CMB signal.[24]

Through the thermal Sunyaev–Zel’dovich effect, energetic electrons in hot gas transfer energy to some passing CMB photons and change the spectrum. The same background that records the early universe can therefore illuminate the later gas in galaxy clusters.[25]

The Hubble tension is a comparison between methods

CMB-based ΛCDM analyses and some local distance-ladder measurements infer different present expansion rates. For example, the 2022 SH0ES Cepheid–supernova analysis reported H₀ = 73.04 ± 1.04 km s⁻¹ Mpc⁻¹, higher than the Planck baseline inference near 67.4. These are historical reference results, not a universal summary of every current measurement.[26]

The discrepancy remains under investigation in 2026. Researchers examine calibration, distance indicators, and cosmological assumptions, and different local methods do not all return the same answer. Changing the late expansion history alone does not automatically reconcile the underlying supernova-calibration comparison.[27]

Unusual large-scale patterns require statistical care

Planck confirmed several previously discussed temperature anomalies, including unusual large-angle features. Its polarization tests did not establish an unambiguous matching cosmological anomaly. A cold spot or alignment can motivate investigation without demonstrating a special cosmic direction, exotic topology, or new physics.[28]

Even a noiseless experiment would face cosmic variance: we observe one sky, with only a limited number of independent large-scale patterns. Better instruments reduce measurement errors, but cannot supply additional realizations of our observable universe. This limitation matters especially when judging rare-looking features on the largest scales.[24]

Primordial gravitational waves remain a search

A combined CMB analysis revised in June 2026 reports r < 0.034 at 95% confidence, where r compares primordial tensor and scalar power. It combines Planck, ACT, SPT, and BICEP/Keck, assumes ΛCDM + r with the single-field tensor consistency relation, and uses k_* = 0.05 Mpc⁻¹. This is an upper limit, not a detection. A null result constrains models predicting stronger tensors while leaving lower-amplitude possibilities open.[29]

ACT’s DR6 measurements extend detailed temperature and E-mode analysis to small angular scales. Combined with other experiments, such data test whether a single cosmological account fits multiple instruments, frequency ranges, and statistical patterns. Published data remain useful long after the observations themselves end.[30]

Mission status checked: 8 September 2026. Development targets describe plans; observing programs and data releases describe different stages of scientific progress.

The strongest advances will connect several measurements: the early plasma, the later distribution of matter, and the expansion history between them. Each new map helps test whether those pieces still describe the same universe.

Cosmology and the Universe’s Large-Scale Structure

1. Cosmic Inflation: Theory and Evidence
2. The Cosmic Web: Filaments, Voids, and Superclusters
3. The Cosmic Microwave Background’s Detailed Structure · You are here
4. Baryon Acoustic Oscillations
5. Redshift Surveys and Mapping the Universe
6. Gravitational Lensing: A Natural Cosmic Telescope
7. Measuring the Hubble Constant: The Tension
8. Dark Energy Surveys
9. Anisotropies and Inhomogeneities
10. Current Debates and Outstanding Questions