Module 7: The cosmic microwave background · Lesson 7.7 · about 15 min
CMB polarisation
A few percent of the CMB's light is polarised: it was scattered off electrons that saw slightly more light from one direction than another. How that happens, what E-modes and B-modes are, and how polarisation tells us when the first stars switched on.
Module 7 · about 15 minutes. Read first: 7.6 WMAP and Planck. So far we have measured only how bright the CMB is in each direction. Light also has a direction of wiggle, its polarisation, and the CMB’s carries a second, independent record of the early universe. By the end you will be able to explain how the CMB was polarised and what E-modes and B-modes are.
1. Light with a direction
Light is a wave of electric and magnetic fields that wiggle across its direction of travel. In ordinary light the wiggles point every which way; in polarised light they prefer one direction. Light scattered off something is often polarised: that is why polarising sunglasses cut the glare from a lake or a road, and why the blue sky looks darker through them in some directions.
When light hits a free electron, it shakes the electron across the light’s direction of travel, and the shaking electron sends out new light (this is Thomson scattering). Light the electron sends towards you is polarised along the direction it shook.
2. How the CMB got polarised
If an electron is lit equally from all sides, the shaking in every direction balances and the scattered light is unpolarised. To get net polarisation, the light around the electron must be uneven in a special way: a little brighter from two opposite sides than from the two at right angles. This pattern is called a quadrupole.
Where did the quadrupoles come from? Only during the brief moment of last scattering, when photons could travel a short way between bounces, could an electron “see” light from slightly different places, and the fluid’s flows into and out of wells made that light uneven. Because the moment was brief, the polarisation is small: about 5% of the temperature ripples, a few microkelvin at most. It is the faintest signal cosmologists have ever mapped.
Because it comes from the fluid’s motion, polarisation peaks where the temperature has troughs: a wave caught halfway through its swing has no temperature extreme but its greatest speed. Seeing the polarisation peaks fall between the temperature peaks was a strong confirmation of the sound-wave picture.
3. E-modes and B-modes
Any pattern of polarisation lines on the sky can be split into two kinds, named after electric and magnetic fields:
- E-modes: lines that run along or across the direction to a spot, in rings or spokes, mirror-symmetric. The sound waves make only these.
- B-modes: lines with a handed swirl, which look different in a mirror. Density waves cannot make them at all.
That is what makes B-modes exciting. They are made in two ways. Gravitational lensing by all the matter between us and the CMB twists some E-modes into B-modes on small scales; this was detected in 2013 by the South Pole Telescope. And gravitational waves from the very first instant would stretch space in a way that makes B-modes on degree scales. Finding those would be direct evidence of inflation. They have not been found yet, and a famous false alarm in 2014 taught how hard the search is (Lesson 11.4). Polarised glow from dust in our own galaxy is the main trap.
4. A second scattering: dating the first stars
Within the first billion years, the first stars and galaxies re-ionised the universe’s hydrogen, freeing electrons again (Lesson 10.5). By then the universe was much thinner, so only a few CMB photons scattered a second time. Those few picked up polarisation on very large angles on the sky. The chance a photon passed through untouched is e−τ:
- τ
- the optical depth to reionisation: how foggy the universe has been since the first stars, counting free electrons along the way
- e−τ
- the fraction of CMB photons that reach us without scattering again
- fscattered
- the fraction that were scattered: about 5% for Planck's value
- Physical meaning
- A bigger τ means the universe was re-ionised earlier. Planck's τ puts the halfway point of reionisation at a redshift of about 7.7, some 700 million years after the start. The scattering also dims the ripples by a factor e−2τ in power, so τ must be known to read the ripples' true strength (Lesson 7.8).
From a prediction to a map. Martin Rees predicted in 1968 that the CMB should be polarised. It took 34 years to see.
- 1968 Rees predicts CMB polarisation
- 1997 Two teams (Kamionkowski, Kosowsky and Stebbins; Seljak and Zaldarriaga) show how to split it into E and B, and that B-modes would reveal gravitational waves
- 2002 DASI, at the South Pole, detects E-mode polarisation
- 2003 WMAP sees large-angle polarisation from reionisation
- 2013 The South Pole Telescope detects lensing B-modes
- 2018 Planck: τ = 0.054; E-modes measured across the whole sky
“A B-mode signal means gravitational waves from inflation.” B-modes are also made by lensing, which is now measured routinely, and polarised dust in our galaxy can mimic them. Only a degree-scale B-mode that survives careful removal of dust, at several frequencies, would count. The 2014 BICEP2 claim fell exactly into the dust trap.
- Thomson scattering of light with a quadrupole pattern leaves the scattered light partly polarised; the CMB is polarised by a few percent of its ripples, a few µK.
- E-modes (rings and spokes) come from the sound waves; they peak between the temperature peaks, as the fluid’s motion predicts.
- B-modes (swirls) come from lensing (detected 2013) and possibly primordial gravitational waves (not yet found).
- Large-angle polarisation from reionisation gives τ = 0.054: about 5% of CMB photons scattered again, reionisation halfway at z ≈ 7.7.
Check yourself
Sources
- M. J. Rees (1968), Polarization and spectrum of the primeval radiation in an anisotropic universe, ApJ 153, L1 https://doi.org/10.1086/180208
- M. Kamionkowski, A. Kosowsky and A. Stebbins (1997), A probe of primordial gravity waves and vorticity, Physical Review Letters 78, 2058 https://doi.org/10.1103/PhysRevLett.78.2058
- U. Seljak and M. Zaldarriaga (1997), Signature of gravity waves in the polarization of the microwave background, Physical Review Letters 78, 2054 https://doi.org/10.1103/PhysRevLett.78.2054
- J. M. Kovac et al. (2002), Detection of polarization in the cosmic microwave background using DASI, Nature 420, 772 https://doi.org/10.1038/nature01269
- D. Hanson et al. (2013), Detection of B-mode polarization in the cosmic microwave background with data from the South Pole Telescope, Physical Review Letters 111, 141301 https://doi.org/10.1103/PhysRevLett.111.141301
- W. Hu and M. White (1997), A CMB polarization primer, New Astronomy 2, 323 https://doi.org/10.1016/S1384-1076(97)00022-5
- Planck Collaboration (2020), Planck 2018 results. VI. Cosmological parameters, A&A 641, A6 https://doi.org/10.1051/0004-6361/201833910
Your ticks stay in this browser. Save them as a file on the course home.
Updated 2026-11-22. Numbers are rounded for teaching; measured values come from the sources listed in each lesson. Simulations run in your browser; nothing you type is sent anywhere.