Cosmic microwave background

• 1. Introduction

  ›

  Point a standard optical telescope at the gaps between stars and galaxies, and the cosmic microwave background stays hidden, because that space looks almost completely dark. A sufficiently sensitive radio telescope tells a different story. It picks up a faint glow, almost perfectly even, that belongs to no star, galaxy, or other object. The glow is strongest in the microwave part of the electromagnetic spectrum, which gives it its name. Some researchers call it relic radiation. Its energy density exceeds that of every photon the stars have emitted across the entire history of the universe. Photons from this background outnumber all other photons by a factor of 400 to 1. If the universe had not expanded and cooled this radiation, the night sky would shine as brightly as the Sun. So where does this light come from, and why does it reach us from every direction at once? Who first guessed that it might exist, and why did others who brushed against it fail to recognize it? And what could tiny flaws in such a smooth glow reveal about the earliest moments of everything?

  1113493

• 2. An Unexplained Excess

  ›

  At Crawford Hill, a Bell Telephone Laboratories site in Holmdel Township, New Jersey, stood a large antenna built in 1959. It was made to support Project Echo, the National Aeronautics and Space Administration's passive communications satellites. Those satellites were huge Earth-orbiting balloons of aluminized plastic, and they bounced radio signals from one point on Earth to another. By 1964 the antenna served a Dicke radiometer that Arno Penzias and Robert Woodrow Wilson had built. They intended it for radio astronomy and satellite communication experiments.

  On the 20th of May 1964, Penzias and Wilson made a first measurement that clearly showed the presence of the microwave background. Their instrument registered an excess antenna temperature of 4.2 K, and they could not account for it.

  At Princeton University, David Todd Wilkinson and Peter Roll, colleagues of Robert H. Dicke, were constructing their own Dicke radiometer that same year. Their goal was to measure the cosmic microwave background. After a telephone call from Crawford Hill, Dicke said, "Boys, we've been scooped." A meeting between the Princeton and Crawford Hill groups confirmed that the antenna temperature really came from the microwave background. Penzias and Wilson received the 1978 Nobel Prize in Physics for their discovery.

  In the spring of 1964, Soviet astrophysicists A. G. Doroshkevich and Igor Novikov had already published a brief paper naming this radiation as something detectable. Penzias and Wilson put its temperature at approximately 3 K. Dicke, James Peebles, Roll and Wilkinson read the signal as a signature of the Big Bang.

  1819202224

• 3. Forgotten Forecasts

  ›

  As early as 1896, Charles Édouard Guillaume estimated the radiation of the stars at 5-6 K. In 1926, Sir Arthur Eddington worked out an effective temperature of 3.18 degrees absolute for starlight in the galaxy. In the 1930s, Erich Regener calculated an effective temperature of 2.8 K for the cosmic rays of the galaxy. These were figures for starlight and cosmic rays, not for a glow left over from a beginning.

  Georges Lemaître speculated in 1931 that remnants of the early universe might be observable as radiation, though he had cosmic rays in mind. In 1934, Richard C. Tolman showed that the expansion of the universe would cool blackbody radiation while keeping its thermal spectrum.

  The true forecast came in 1948 from Ralph Alpher and Robert Herman, in a correction they prepared for a paper by Alpher's doctoral advisor, George Gamow. They reasoned that if a Big Bang had happened, expansion would have stretched the high-energy radiation of the early universe into the microwave region. They put its temperature at about 5 K. The estimate was slightly off, but the idea was right. Gamow himself had calculated 50 K in 1946, assuming a universe 3 billion years old, without mentioning background radiation. In 1953 he estimated 7 K.

  Several observers came close without knowing it. In 1941, Andrew McKellar detected a rotational temperature of 2.3 K in the interstellar medium, by comparing CN doublet lines that W. S. Adams had measured in a B star. In 1955, Émile Le Roux of the Nançay Radio Observatory surveyed the sky at a wavelength of 33 cm. He reported a near-isotropic background of 3 kelvins, plus or minus 2, but did not recognize its cosmological significance.

  In 1957, Tigran Shmaonov reported a background of 4 plus or minus 3 K, with an intensity that did not depend on time or direction. He was observing at a wavelength of 3.2 cm, and it is now clear that he had seen the cosmic microwave background. In 2005, Alpher was awarded the National Medal of Science for his work in nucleosynthesis and for predicting that the expanding universe leaves background radiation behind.

  1415162526

• 4. Proving It Cosmic

  ›

  Through the late 1960s, the meaning of the signal was a controversial question. Proposed explanations included energy from within the Solar System, from galaxies, from intergalactic plasma, and from multiple extragalactic radio sources.

  Two tests could show that the radiation was truly cosmic. First, its spectrum had to match a thermal, or black body, source. By 1968, a series of measurements of the radiation temperature at higher and lower wavelengths had shown that it did. Second, the radiation had to be isotropic, meaning the same from every direction. That was demonstrated by 1970.

  The discovery of the mid-1960s curtailed interest in alternatives such as the steady state theory. Together with the cosmological redshift-distance relation, the background is regarded as the best available evidence for the Big Bang. Its measurements made the inflationary Big Bang model the Standard Cosmological Model.

  Standard cosmology still has critics. It requires fine-tuning of some free parameters, and different experimental data support different values. It cannot predict the present temperature of the relic radiation, while the steady state model can. Alternative models, however, have only offered after-the-fact explanations of existing observations. They have still played a historic role, supplying ideas for the standard explanation and challenges to it.

  In 1990, the FIRAS instrument on the Cosmic Background Explorer satellite, known as COBE, measured the spectrum with exquisite precision. It found a nearly perfect black body at 2.73 K. The background has since become the most precisely measured black body spectrum in nature.

  21064451

• 5. The Fog Lifts

  ›

  Roughly 380,000 years after the Big Bang, the universe had cooled to about 3,000 K, and its character changed. Until then it was filled with an opaque fog of dense, hot plasma made of sub-atomic particles. Photons kept scattering off free electrons in that fog, an effect called Thomson scattering. As expansion cooled the plasma, protons and electrons combined into neutral atoms, mostly hydrogen. These atoms could not scatter radiation that way, so the universe became nearly transparent. This moment is called recombination, or decoupling, and it released photons to travel freely through space.

  The team behind the Wilkinson Microwave Anisotropy Probe, or WMAP, measured how long the change took. Its first-year results put the most likely moment of last scattering at 372,000 years. Decoupling stretched over roughly 115,000 years, so by the time it finished, the universe was about 487,000 years old.

  Since then, the temperature of that light has dropped by an average factor of 1,089 as the universe expanded. Today it stands at about 2.725 K, and it will keep falling. The photons we receive arrive from the surface of last scattering, a spherical shell of locations at just the right distance for light from decoupling to be reaching us now.

  Cosmologists semi-humorously call the interval between this release of light and the first stars the Dark Age. Observations of galaxies show that most of the volume of the intergalactic medium is ionized, because few hydrogen absorption lines appear. That implies a period of reionization. In an ionized universe, free electrons scatter the background photons, and WMAP observed two resulting effects. Small scale anisotropies are erased, much as details blur when looking through fog, and polarization appears on large angular scales. Together they suggest the universe was ionized at a redshift around 10.

  What did the ionizing is still debated. Candidates include light from the very first population III stars, supernovae from those stars, and radiation from the accretion disks of massive black holes.

  By 1969, R. A. Sunyaev and Yakov Zel'dovich were studying one further change to the light on its journey. In the Sunyaev-Zel'dovich effect, a cloud of high-energy electrons scatters the radiation and passes some of its energy to the background photons. The Sachs-Wolfe effect works differently, gravitationally redshifting or blueshifting photons as gravitational fields change. In 1983, researchers from the Cambridge Radio Astronomy Group and the Owens Valley Radio Observatory first detected the Sunyaev-Zel'dovich effect from clusters of galaxies.

  634051304

• 6. Reading the Ripples

  ›

  Across the whole sky, the background is isotropic to roughly one part in 25,000. The root mean square variations are just over 100 microkelvin, once a dipole anisotropy is subtracted. That dipole is a Doppler shift caused by the Sun moving at 369.82 plus or minus 0.11 kilometers per second toward the constellation Crater, near its boundary with Leo. It carries no primordial information, and it is now used to calibrate mapping studies.

  In the 1970s, Harrison, Peebles and Yu, and Zel'dovich realized that quantum inhomogeneities in the early universe would leave temperature differences of one part in ten thousand or one in a hundred thousand. Rashid Sunyaev calculated the observable imprint they would make. A lull followed, partly because measuring the background at high precision was so difficult. Ground-based experiments set ever tighter limits during the 1980s. RELIKT-1, a Soviet experiment aboard the Prognoz 9 satellite launched on the 1st of July 1983, gave the first upper limits on large-scale anisotropy.

  Alan Guth's proposal of cosmic inflation in the 1980s explained the large-scale evenness of the glow. His theory of rapid spatial expansion allowed distant regions to be causally connected just before the epoch of last scattering. Detailed predictions like this encouraged larger and more ambitious experiments.

  NASA's Cosmic Background Explorer orbited Earth from 1989 to 1996. Its Differential Microwave Radiometer clearly confirmed the primary anisotropy, and the team published its findings in 1992. In January 1992, scientists analysing RELIKT-1 data had reported anisotropy at the Moscow astrophysical seminar. In 2006, COBE's principal investigators George Smoot and John Mather received the Nobel Prize in Physics.

  The ripples form a pattern of peaks and valleys that varies with angular scale. They arise from a contest in the early plasma. The pressure of photons tends to erase irregularities, while the gravitational attraction of the baryons pulls matter into overdensities. The first peak determines the overall curvature of the universe. The second peak details the density of normal matter, and the third details the density of dark matter.

  COBE lacked the resolution to see the first peak. The MAT/TOCO experiment tentatively detected it, and the BOOMERanG and MAXIMA experiments confirmed the result. By 2000, BOOMERanG reported that the strongest fluctuations occur at scales of approximately one angular degree. These measurements showed that the geometry of the universe is approximately flat. They also ruled out cosmic strings as a major part of cosmic structure formation and favored cosmic inflation.

  In June 2001, NASA launched WMAP to measure the large-scale anisotropies over the full sky. It used rapid-switching radiometers at five frequencies to minimize noise that did not come from the sky. Its data came out in five installments, the last being the nine-year summary. The results broadly fit Lambda CDM models based on 6 free parameters, within Big Bang cosmology with cosmic inflation.

  The European Space Agency launched the Planck Surveyor in May 2009, and it operated until October 2013. It used both HEMT radiometers and bolometer technology, and it measured the background at a smaller scale than WMAP. On the 21st of March 2013, the Planck team released its all-sky map. The map suggested the universe is slightly older than researchers expected. Its temperature fluctuations were imprinted when the cosmos was about 370,000 years old, and they reflect ripples that arose as early as the first nonillionth of a second.

  Based on the 2013 data, the universe contains 4.9% ordinary matter, 26.8% dark matter and 68.3% dark energy. On the 5th of February 2015, new Planck data put the age of the universe at 13.799 billion years and the Hubble constant at 67.74. The final Planck data and maps arrived in 2018, with improved measurements of polarization on large scales.

  313238406

• 7. Polarization and a Retraction

  ›

  A few microkelvin: that is the level at which the background is polarized. The polarization comes in two types, called E-mode, or gradient mode, and B-mode, or curl mode. The names echo electrostatics, where the electric field has a vanishing curl and the magnetic field has a vanishing divergence. E-modes arise from Thomson scattering in a heterogeneous plasma. They are a factor of 10 weaker than the temperature anisotropy. The Degree Angular Scale Interferometer first saw them in 2002, and the Cosmic Background Imager then produced the first E-mode spectrum, which was out of phase with the temperature spectrum.

  B-modes are expected to be weaker than E-modes by about an order of magnitude. Standard scalar perturbations do not produce them. Gravitational waves from cosmic inflation, shortly after the Big Bang, should. Models of slow-roll inflation predict a specific B-mode pattern, and its strength could confirm some models of inflation and exclude others.

  On the 17th of March 2014, the BICEP2 collaboration announced the detection of inflationary gravitational waves in the B-mode power spectrum. Confirmed, it would have been clear experimental evidence for the theory of inflation. On the 19th of June 2014, lowered confidence in the finding was reported. On the 30th of January 2015, the same astronomers withdrew the claim. Using combined BICEP2 and Planck data, the European Space Agency announced that the signal could be entirely attributed to dust in the Milky Way. Later reanalysis that compensated for foreground dust gave limits in agreement with Lambda-CDM models.

  A second kind of B-mode does exist. Gravitational lensing of the stronger E-modes also produces B-mode polarization, and the South Pole Telescope discovered this type in 2013 with help from the Herschel Space Observatory. In October 2014, the POLARBEAR experiment published a B-mode measurement at 150 GHz. It looks at a smaller patch of sky than BICEP2 and is less affected by dust. Its team reported a 97.2% confidence level that the signal was of cosmological origin.

  Spectral distortions, tiny departures from the black body law, are the next target. Researchers hope to make a first measurement of them within the forthcoming decades, because they should hold a wealth of information about the primordial universe.

  6667767377

• 8. Anomalies and the Far Future

  ›

  Even in the COBE map, the quadrupole, a pattern at the largest angular scales, showed a lower amplitude than Big Bang predictions called for. As WMAP delivered more precise data, claims of anomalies multiplied. They included very large scale anisotropies, odd alignments, and non-Gaussian distributions. The longest-running is the low-multipole controversy. The quadrupole and octupole modes appear aligned with each other and with both the ecliptic plane and the equinoxes. Some groups suggest this is a signature of quantum corrections or new physics at the greatest observable scales. Others suspect systematic errors in the data.

  A full Bayesian analysis of the WMAP power spectrum found the quadrupole prediction of Lambda-CDM cosmology consistent with the data at the 10% level. It also found the observed octupole unremarkable. Carefully accounting for the way foregrounds were removed from the full sky map reduced the significance of the alignment by about 5%.

  Planck, much more sensitive than WMAP and with a larger angular resolution, recorded the same anomaly. That appears to rule out instrumental error, though not foreground contamination. WMAP chief scientist Charles L. Bennett suggested coincidence and human psychology: "I do think there is a bit of a psychological effect; people want to find unusual things."

  Measurements of the density of quasars, based on Wide-field Infrared Survey Explorer data, find a dipole significantly different from the one extracted from the background's anisotropy. That difference conflicts with the cosmological principle.

  Assuming the universe keeps expanding and avoids a Big Crunch or a Big Rip, the background will keep redshifting until it can no longer be detected. Starlight will supersede it first. Later, background radiation fields from proton decay, the evaporation of black holes, positronium decay, or Unruh-effect radiation associated with the particle horizon may take over.

  85981008681