The Cosmic Microwave Background (CMB)

Before recombination, photons frequently scattered from free electrons. Cooling allowed neutral atoms to form, reducing scattering and extending photon paths. Most CMB light reaching us last scattered around 380,000 years after the Big Bang, at roughly 3,000 K.[1]

Recombination means atom formation; decoupling describes weakened scattering. Last scattering identifies the final scattering event along a photon’s observed journey.[1]

During expansion, freely propagating radiation stretches to longer wavelengths. The background’s temperature falls inversely with the cosmic scale factor.[3]

A widely used precision estimate is 2.72548 ± 0.00057 K, from a 2009 combined analysis dominated by COBE’s FIRAS measurements. For most explanations, 2.725 K is enough. This is the radiation’s mean temperature, not a claim that every cloud, star, or patch of empty space has that temperature.[4]

COBE/FIRAS found that the observed spectrum follows a blackbody curve with extraordinary accuracy. This strongly supports a hot, thermalized early universe. Small allowed departures constrain later energy release; agreement with a blackbody does not mean that no photon ever scattered again or that absolutely no later energy was injected.[5]

In the 1940s, work on the hot early universe led Ralph Alpher and Robert Herman to predict a surviving, cooled radiation background. Their calculation emerged from the wider effort associated with George Gamow to understand early cosmic conditions and element formation.[6]

In 1964–1965, Arno Penzias and Robert Wilson found an unexplained microwave signal using a Bell Labs horn antenna. Their 1965 paper reported an excess at 4.08 GHz and pointed to a companion paper by the Princeton group for a cosmological interpretation. This discovery was not yet a measurement of the complete blackbody spectrum.[7]

Each mission improved a different part of the measurement. Greater angular resolution, more frequency bands, better calibration, and lower noise all matter; no single specification captures the quality of a cosmic map.

The largest familiar directional variation is the dipole: one side of the sky appears slightly warmer and the opposite side cooler, mainly because of our motion relative to the CMB. Researchers account for this before studying the much smaller primordial pattern.[10]

The remaining temperature differences are typically of order one part in 100,000. COBE’s early maps detected variations of tens of microkelvins, with the exact amplitude depending on angular smoothing and the sky region analyzed. These are minute departures from the background’s mean temperature.[8]

The familiar red and blue patches are colors assigned to data. They do not show red-hot and blue-cold objects. Nor does every hotter patch translate directly into a denser patch of matter: intrinsic radiation temperature, plasma motion, gravity at last scattering, and changes along the light’s journey all contribute.[2]

Before decoupling, photons and ordinary matter behaved as a coupled fluid. Gravity encouraged compression while photon pressure resisted it. The resulting oscillations affected regions of many sizes. At last scattering, different wavelengths were caught at different stages of compression and rarefaction.[12]

An angular power spectrum summarizes how much temperature variation occurs on different angular scales. Its acoustic peaks encode this oscillatory history. They are statistical features of the sky pattern, not a series of separate explosions.[12]

• Peak positions connect a physical acoustic scale with its apparent size on the sky.[12]
• Relative peak heights respond to the ordinary-matter content and to gravitational driving associated with the matter–radiation balance.[12]
• The damping tail records how photon diffusion smoothed sufficiently small-scale fluctuations.[12]

The peak pattern is more informative than any one peak. Its interpretation requires a model of early physics and cosmic distances; the first peak alone is not a standalone proof of spatial flatness.[13]

Light is an electromagnetic wave. Linear polarization describes a preferred orientation of its electric-field oscillations. Scattering can generate it when an electron is illuminated by radiation with a suitable directional imbalance, called a quadrupole pattern.[14]

Across the sky, researchers divide polarization into E-mode and B-mode patterns. These describe spatial geometry, not separate types of photons or direct maps of cosmic electric and magnetic fields. At linear order, primordial density fluctuations generate E modes; primordial gravitational waves can also generate B modes.[14]

DASI first detected CMB polarization in 2002, including the predicted E-mode signal.[15] WMAP, Planck, and other instruments later measured polarization with increasing detail.[9], [10]

B modes also arise when gravitational lensing remaps existing E-mode polarization. Those lensing B modes have been detected. They must be distinguished from a possible primordial gravitational-wave contribution.[16]

As of September 2026, no primordial B-mode signal has been confirmed. Galactic dust and synchrotron emission complicate the search, while lensing adds a genuine cosmological contribution that researchers model or remove statistically. A B-mode pattern by itself does not identify inflation.[17], [18]

Researchers calculate predicted temperature, polarization, and lensing patterns, then compare them with observations. The standard ΛCDM model combines ordinary matter, cold dark matter, and a cosmological constant. Its ability to fit several kinds of CMB measurement with a consistent parameter set is a major success.[13]

Within that framework, the CMB constrains matter densities, the initial fluctuation spectrum, and cosmic distances. Planck’s baseline analysis inferred H₀ = 67.4 ± 0.5 km/s/Mpc. This is a model-dependent inference from early-universe observations, not a direct local speed measurement.[13]

A 2022 Cepheid–supernova distance-ladder analysis found about 73 km/s/Mpc. Comparing these approaches illustrates the Hubble tension: disagreement can motivate checks of measurements, calibration, and theory, but it does not identify a unique missing ingredient.[20]

CMB patterns also test inflationary ideas. Their nearly scale-invariant primordial spectrum is consistent with many inflation models, yet it does not uniquely establish a particular mechanism. The detailed physics of inflation remains open.[21]

Combining CMB data with galaxy clustering, lensing, and supernova distances helps separate effects that resemble one another in the CMB alone. Tight statements about curvature or dark energy therefore need the accompanying model and dataset assumptions.[13]

Research context: September 2026. The frontier includes finer temperature and polarization measurements, better foreground separation, and stronger tests for primordial gravitational waves.

More sensitive instruments cannot remove every limitation. We observe one cosmic sky, with only a limited number of independent patterns at the largest angular scales. This unavoidable statistical limitation is called cosmic variance; foregrounds and instrument noise add further uncertainty.[2]

For a closer look at acoustic peaks, E and B modes, and cosmological inference, continue to The Cosmic Microwave Background’s Detailed Structure.

The Grand Beginning

1. The Singularity and Moment of Creation
2. Quantum Fluctuations and Inflation
3. Big Bang Nucleosynthesis
4. Matter vs. Antimatter
5. Cooling and the Formation of Fundamental Particles
6. The Cosmic Microwave Background (CMB) · You are here
7. Dark Matter
8. Dark Energy
9. Recombination and the First Atoms
10. The Dark Ages and First Structures
11. Reionization: Ending the Dark Ages