On Monday, March 17, 2014, a press conference was convened by the Harvard-Smithsonian Center for Astrophysics (CfA).1 Four scientists representing the Background Imaging of Cosmic Extragalactic Polarization (BICEP2) experiment made a monumental claim.2

Researchers from the BICEP2 collaboration today announced the first direct evidence for … cosmic inflation. Their data also represent the first images of gravitational waves, or ripples in space-time.3

Their discovery turned on the telltale signs of the inflationary epoch theorized by cosmologists.

These groundbreaking results came from observations by the BICEP2 telescope of the cosmic microwave background—a faint glow left over from the Big Bang. Tiny fluctuations in this afterglow provide clues to conditions in the early universe. For example, small differences in temperature across the sky show where parts of the universe were denser, eventually condensing into galaxies and galactic clusters.4

The BICEP2 findings were greeted with satisfaction by both the scientific press and the mainstream media. “Space Ripples Reveal Big Bang’s Smoking Gun,” proclaimed a headline in The New York Times.5 “Telescope Captures View of Gravitational Waves,” Nature affirmed.6 “Gravitational Waves from Big Bang Detected,” echoed Scientific American.7

The Invisible Worm

Just after the Big Bang, the cosmos is thought to have grown even faster than the speed of light. This idea of exponential expansion is the central claim of the inflationary hypothesis developed by Alan Guth and Andrei Linde.8 Quantum fluctuations of the inflationary field are thought to have seeded the formation of galaxies, triggering fluctuations in the density of radiation that in turn determined peaks and valleys in the distribution of matter.9

Forced for almost 380,000 years to interact within the hot plasma constituting the early universe, photons were finally freed when expansion cooled the plasma and allowed protons and electrons to bond into hydrogen atoms. Residual electromagnetic radiation, the cosmic microwave background (CMB), has permeated the entire cosmos as an afterglow ever since, with expansion increasingly stretching its wavelength into microwaves. The search for the characteristic imprints of inflation in the CMB have been the focus of research seeking support for the theory of inflation.

After 14 billion years of expansion, the CMB has cooled to 2.7 Kelvin (i.e. -270ºC). This temperature is almost exactly the same everywhere. Minuscule variations are present on the order of one part in 100,000. By measuring anisotropic changes in the CMB, physicists are able to ascertain the behavior of nature at energy levels that no particle accelerator will ever be able to probe.

At the very beginning of the universe, fluctuations of space and time at the tiniest scales were stretched by inflation into gravitational waves. As a result of their propagation, space and time was pulled in one of the directions perpendicular to their motion, and squeezed in the other. When light encountered this alternation of bumps and depressions, or throats and valleys, its photons acquired a peculiar curly twist, or polarization. This is the B-mode polarization the BICEP2 experiment was searching for in the CMB photons.

When physicists examine the CMB light, they are observing the universe at the largest scale. The observed backdrop is not very elaborate. It is almost as plain as possible, with few distinguishing features other than small anisotropies and their global pattern. The universe as a whole is homogenous and smooth, its global space-time curvature flat.

Such simplicity is a real challenge in cosmology.

Inflationary theories claim to deliver the necessary smoothing and stretching. However appealing these theories might be, neither beauty nor elegance should spare them the rigor of experimental testing.

The BICEP2 team was convinced that they had found the B-mode polarizations in the CMB, the telltale peculiar curly threads in the background tapestry of light. The presence of these patterns would have been a strong reason to favor inflation as the best current explanation of what happened in the first fraction of a second.

In Chickentown

Despite some whispers beforehand, the announcement of BICEP2’s findings in March 2014 came as a complete surprise to most observers. The team had maintained strict secrecy during preparations for the release of their results.10

Fireworks followed the BICEP2 announcement. A few cosmologists and experimentalists managed to contain their enthusiasm.11 Their voices were largely drowned out amidst the hubbub.

Within days of the announcement, the media had already begun speculating on the likelihood of Nobel prizes being awarded to the pioneers of inflationary theory. If BICEP2’s findings were confirmed, The Guardian suggested, “the 2015 Nobel prize committee will have a tough choice to make,” apparently because of an embarrassment of riches.12 BBC News went further: “Assuming the BICEP2 results are confirmed, a Nobel Prize seems assured.”13

The excitement surrounding the BICEP2 results also became an unexpected social media phenomenon. A widely circulated YouTube video documented an impromptu visit from a member of the BICEP2 collaboration to Linde’s home to personally deliver the news.14 BICEP2 had found evidence for inflation. As the video draws to a close, a bottle of champagne is opened and glasses are raised.

For the members of the BICEP team, it must have seemed only a matter of time until the initial acclaim led to a formal acknowledgment of their findings. This was the case when the discovery of the Higgs boson was announced by the European Center for Nuclear Research (CERN) in July 2012. A year later, Peter Higgs and François Englert were awarded the Nobel Prize in Physics.

There was, however, one major difference between the BICEP2 and CERN announcements. CERN was able to make its claim because two independent experiments had spotted the clearest hints of the elusive boson.15

At roughly the same time as the public announcement, the BICEP2 team submitted their work for peer review.

What a mistake.

The Evidence from Planck

The BICEP2 team was far from being the only group observing the CMB. Launched in March 2009, the European Space Agency’s (ESA) Planck satellite had also been observing and mapping the anisotropies of the CMB—more broadly so than BICEP2.16

A paper published by the Planck team in September 2014 containing their latest findings offered no support for the results claimed by BICEP2. On the contrary, extrapolation of the Planck data to the BICEP2 observation frequency window yielded a dust contribution of the same magnitude as that reported by BICEP2. This, according to the Planck team, “highlights the need for assessment of the polarized dust signal even in the cleanest windows of the sky.”17

While the signal observed by BICEP2 had, indeed, been real, it was a signal that did not originate from the dawn of the universe.

As CMB photons travel through space, they are absorbed and re-emitted by intergalactic dust. CMB light is twisted in B-modes, the same swirly polarization induced by inflationary gravitational waves. The CMB constitutes almost all the light in the universe, but it is also faint. Its signal may well be contaminated. The BICEP2 team was confident that this issue had been addressed.

“The key question,” says Daniel Eisenstein, an astrophysicist at the CfA, “is whether there could be a foreground that masquerades like this signal”. But the team has all but ruled out that possibility, he says. First, the researchers were careful to point BICEP2—an array of 512 superconducting microwave detectors—at the Southern Hole, a patch of sky that is known to contain only tiny amounts of such emissions. They also compared their data with those taken by an earlier experiment, BICEP1, and showed that a dust-generated signal would have had a different colour and spectrum.18

The signal detected by BICEP2 was only as trustworthy as their assumptions about foregrounds. These were incomplete because no one had yet addressed the question of whether the single observation frequency chosen by the BICEP2 team was indeed immune to contaminants. There remained a possibility that a sizeable amount of the B-mode polarization, which BICEP2 believed to be of primordial origin, had instead been imprinted on light by its interactions with intergalactic dust.

During the first half of 2014, the Planck team was occupied in analyzing the data captured by their satellite prior to its scheduled decommissioning.19

They would soon be in a position to know the strength of B-mode contamination.

Having had no access to Planck’s results before they made their announcement, the BICEP2 team was forced to extract data from slides the Planck team had presented during a conference talk the previous year.20

Instead of recognizing the incompleteness of their claims and the risk this posed to the validity of their results, the BICEP2 team persevered. They had made a huge claim before submitting their results to external scrutiny.

Following the release of their dust maps at the end of September, the Planck team began working with the BICEP2 team.21 Results from the joint analysis were released at the end of January 2015. Although Planck had tracked the dust trail in a different frequency and at a lower specific sensitivity than BICEP2, Planck’s dust signal was strongest in precisely the same region of the sky that BICEP2 had been observing.22

The verdict was clear. BICEP2’s claim to have found evidence of primordial gravitational waves could not be supported. There remains a sliver of hope that a genuine primordial signal might yet be found in the BICEP2 data. However, if such a signal were found, it would be much smaller than that originally claimed by the BICEP2 team.

Gravitational Waves and the Inflation Debate

The dilution of BICEP2’s findings was a chastening experience for the scientists involved. It was also a blow to the theory of inflation. Ever since the BICEP2 results were announced, there has been renewed debate about inflation.23 It is not the existence of gravitational waves that is being questioned, but the validity of inflation as a theory to explain the infant universe.

Some inflationary models predict primordial gravitational waves, hence the experimental efforts to find them.24 The search for this type of gravitational wave is focused on an indirect effect, namely the imprint left behind in the CMB polarization. If this imprint cannot be detected, the inflationary models will have been disfavored experimentally. It does not mean that gravitational waves do not exist.

Gravitational waves are disturbances in the fabric of space and time. In the case of inflation, it is thought that they originate from exponentially stretching the quantum fluctuations of space-time itself.25 Gravitational waves are thought to be produced by energetic events such as supernovae or binary systems of black holes and neutron stars spiraling around each other. Gravitational waves from these phenomena have not yet been directly observed, but strong indirect evidence for their existence has been found. These gravitational waves have a very different theoretical basis from primordial inflationary waves.

Inflationary theories seem to provide the missing link between quantum mechanics and cosmology. This link spans an energy range that is many orders of magnitude wide and remains largely unprobed. Uncertainty is increased by the fact that we still lack a theory of quantum gravity: a necessary framework when tackling the most fundamental scales of energy, space, and time. That properly moving astrophysical masses generate gravitational waves follows directly from Albert Einstein’s theory of General Relativity. This is our best current explanation of gravity.26

A particularly revealing test of General Relativity concerns a system of two neutron stars, one of which is a pulsar. The binary pulsar discovered by Russell Hulse and Joseph Taylor in 1974 provided observational evidence that gravity propagates at the speed of light.27 While rotating on itself, the magnetic field of such a pulsating star produces a radio signal, a lighthouse, of sorts, that can be detected from earth if the bodies are favorably aligned.28 By means of this signal, we can establish that the two stars are orbiting a common center of mass at high speed and in a compact configuration.

General relativity also predicts that a binary system could not be stationary. In order to have a sizable effect, astrophysical bodies need to move fast—at least a thousand times faster than they do in the solar system—and to be very compact.

Other systems have been found that provide further indications of the existence of gravitational waves.29 This evidence is indirect in two ways. Departures from stationary orbital motion of one, both, bodies is detectable in binary systems. Mechanical energy is lost to the production of gravitational waves. Furthermore, this evidence is conveyed mainly by electromagnetic waves, not only by radio signals.30

This situation is about to change. In the years ahead, gravitational wave antennae should begin detecting changes in the relative distance of the instruments’ test masses, which would be evidence of a passing gravitational wave. Not only will this detection be a direct confirmation of the predicted phenomena, but it will also unveil higher order effects than the binary pulsar laboratory could.31

Inflation cannot be regarded with the same degree of confidence.

Theoretical foundations are still shaky because we lack a complete framework that encompasses gravity and quantum mechanics; its desired features are intertwined with specific initial conditions; and its models seem to become ever more numerous and baroque as new data reveals a more stringently simple universe.32

If a simple universe demands a simple theory, it might be worthwhile to consider theoretical schemes competing with inflation. Paul Steinhardt and Neil Turok have proposed that the Big Bang does not represent the beginning of time, but rather the start of one among many phases of expansion, each of which is followed by a period of contraction. It is this combination of expansion and contraction that “produces the homogeneity, flatness, density fluctuations and energy needed to begin the next cycle.”33

A central concern in the discussion evaluating alternative explanations for the origin of the universe should be their testability and the distinctive character of their predictions. However, the ultimate fate of BICEP2’s results has had little effect on the unflinching conviction of inflation’s supporters. This, in itself, would seem to point to a deeper problem in the field.34


A year after announcing their findings, the BICEP2 team has acknowledged its mistakes and downgraded its claims. Purified and strengthened by the fire of online and offline peer review, they are now better prepared to undertake a third campaign of observations at the South Pole, BICEP3.35

They will not be alone in their search for B-mode polarization in the CMB. Other current experimental projects include the Atacama B-Mode Search led by Princeton University; the POLARBEAR experiment led by the University of California, Berkeley; the high-altitude balloon-borne E and B Experiment led by the University of Minnesota; and the Cosmology Large Angular Scale Surveyor led by Johns Hopkins University.

  1. CfA Press, “BICEP2 Press Conference - March 17, 2014,” YouTube video, March 18, 2014. 
  2. Peter Ade et al., “BICEP2 I: Detection Of B-mode Polarization at Degree Angular Scales,” (2014): 13, doi: arXiv:1403.3985v1. This is the first, non-peer-reviewed, version of the paper published in March 2014 claiming the primordial nature of the detected B-modes. The initial version of the paper concludes with the following statement: “The long search for tensor B-modes is apparently over, and a new era of B-mode cosmology has begun.” A second, peer-reviewed, version of the paper with methodological corrections and considerably more modest claims was published in June 2014. Peter Ade et al., “BICEP2 I: Detection of B-mode Polarization at Degree Angular Scales,” Physical Review Letters 112, no. 241101 (2014): 1–25. 
  3. Harvard Smithsonian Center for Astrophysics, “First Direct Evidence of Cosmic Inflation,” March 17, 2014 
  4. Harvard Smithsonian Center for Astrophysics, “First Direct Evidence of Cosmic Inflation,” March 17, 2014. 
  5. Dennis Overbye, “Space Ripples Reveal Big Bang’s Smoking Gun,” The New York Times, March 17, 2014. 
  6. Ron Cowen, “Telescope Captures View of Gravitational Waves,” Nature 507, no. 7492 (2014): 281–83. 
  7. Clara Moskowitz, “Gravitational Waves from Big Bang Detected,” Scientific American, March 17, 2014. 
  8. Although widely credited to Alan Guth, Andrei Linde, and often Paul Steinhardt, the theory of inflation has a number of fathers: Aleksei Starobinsky, “A New Type of Isotropic Cosmological Model Without Singularity,” Physics Letters B91 (1980): 99–102; Alan Guth, “Inflationary Universe: A Possible Solution to the Horizon and Flatness Problems,” Physical Review D 23, no. 2 (1981): 347–56; Andrei Linde, “A New Inflationary Universe Scenario: A Possible Solution of the Horizon, Flatness, Homogeneity, Isotropy, and Primordial Monopole Problems,” Physics Letters B108, no. 6 (1982): 389–393. For an overview see: Viatcheslav Mukhanov, Physical Foundations of Cosmology (Cambridge: Cambridge University Press, 2005). 
  9. Viatcheslav Mukhanov and Gennady Chibisov, “Quantum Fluctuations and a Nonsingular Universe,” Journal of Experimental and Theoretical Physics Letters 33 (1981): 532–35. 
  10. Adam Mann, “How the Biggest Scientific Discovery of the Year Was Kept a Secret,” Wired, March 20, 2014. 
  11. Raphael Flauger, J. Colin Hill, and David Spergel, “Toward an Understanding of Foreground Emission in the BICEP2 Region,” Journal of Cosmology and Astroparticle Physics 1408, no. 039 (2014); Michael Mortonson and Uroš Seljak, “A Joint Analysis of Planck and BICEP2 B Modes Including Dust Polarization Uncertainty,” Journal of Cosmology and Astroparticle Physics 10 (2014): 035. 
  12. Stuart Clark, “Gravitational Waves Give Nobel Prize Committee Another Headache,” The Guardian, March 21, 2014. 
  13. Jonathon Amos, “Cosmic Inflation: ‘Spectacular’ Discovery Hailed,” BBC News, March 17, 2014. 
  14. Stanford University, “Stanford Professor Andrei Linde Celebrates Physics Breakthrough,” YouTube video, March 17, 2014. To date, the video has been viewed nearly 3,000,000 times. 
  15. ATLAS Collaboration, “Observation of a New Particle in the Search for the Standard Model Higgs Boson with the ATLAS Detector at the LHC,” Physics Letters B 716, no. 1 (2012): 1–29; CMS Collaboration, “Observation of a New Boson at a Mass of 125 GeV with the CMS Experiment at the LHC,” Physics Letters B 716, no. 1 (2012): 30–61. 
  16. European Space Agency, “Planck Summary.” 
  17. Planck Collaboration, “Planck Intermediate Results. XXX. The Angular Power Spectrum of Polarized Dust Emission at Intermediate and High Galactic Latitudes,” Astronomy & Astrophysics (2014). 
  18. Ron Cowen, “Telescope Captures View of Gravitational Waves,” Nature 507, no. 7492 (2014): 281–83. 
  19. After four and a half of years of observation and data collection, the Planck satellite reached its final operational orbit on July 3, 2009 was and switched off on October 23, 2013. European Space Agency, “Planck Fact Sheet.” 
  20. This is explicitly stated in the first, non-peer reviewed version of the BICEP2 Collaboration’s paper in the caption for Figure 6:
    Polarized dust foreground projections for our field using various models available in the literature, and two new ones formulated using publically available information from Planck.
    The sources acknowledged in footnotes 33 and 34 are slides from two seminars given by members of the Planck collaboration at an ESA Symposium, held in April, 2013.

    BICEP2 Collaboration, “BICEP2 I: Detection Of B-mode Polarization at Degree Angular Scales,” (2014): 13, doi: arXiv:1403.3985v1.

    One of the two Planck models the BICEP2 team obtained by reverse engineering plots from slides was considered “the best available estimate.” The claim that the detected B-modes were of primordial origin was especially reliant on this particular model. In the second version of the paper, following peer review, the BICEP2 team recognized the uncontrolled uncertainty of this unorthodox process for obtaining reference data and it was eliminated from their analysis, thus mitigating their final claims.

    In common with the adoption of public media by the BICEP2 Collaboration, the behind-the-scenes aspects of these technical issues have been discussed at length in blogs and other online forums, including Facebook. The public character of the debate should not be misconstrued as being uninformed or scientifically inaccurate. On the contrary, many online contributions were an integral part of the peer-review process that the BICEP2 collaboration had previously dodged. See, for example, the following blog posts by physicists Adam Falkowski and Peter Coles: Adam Falkowski, “Is BICEP Wrong?Résonaances, May 12, 2014; Peter Coles, “That BICEP Rumour,” In the Dark, May 14, 2014. 
  21. Jonathon Amos, “Cosmic Inflation: BICEP2 and Planck to Share Data,” BBC News, July 3, 2014; Ian O’Neill, “‘Big Bang’ Scientists to Team-up With Planck Space Telescope?Discovery News, July 3, 2014. 
  22. Peter Ade et al. (BICEP2/Keck and Planck Collaborations), “A Joint Analysis of BICEP2/Keck Array and Planck Data,” Physical Review Letters 114, no. 101301 (2015). 
  23. “Primordial Universe after Planck,” was a symposium held at the Institute for Astrophysics in Paris, in December 2014. Videos from the event have been published online
  24. See the report from the “Dark Energy and CMB” working group for the American Physical Society's Division of Particles and Fields long-term planning exercise (“Snowmass”): K. N. Abazaijan et al., “Inflation Physics from the Cosmic Microwave Background and Large Scale Structure,” (2013) doi: arXiv:1309.5381. 
  25. Aleksei Starobinsky, “Spectrum of Relict Gravitational Radiation and the Early State of the Universe,” Journal of Experimental and Theoretical Physics Letters 30, no. 11 (1979): 682–85. 
  26. Albert Einstein, “The Foundation of the General Theory of Relativity” (in German), Annalen der Physik 49 (1916) 769–822. See also The Collected Papers of Albert Einstein, a newly established and comprehensive archive. 
  27. Russell Hulse and Joe Taylor, “Discovery of a Pulsar in a Binary System,” Journal of Astrophysics 195 (1975) L51–L53; Robert Wagoner, “Test for the Existence of Gravitational Radiation,” Journal of Astrophysics 196 (1975): L63–L65. 
  28. Anthony Hewish et al., “Observation of a Rapidly Pulsating Radio Source,” Nature 217 (1968): 709–13. See also the following material for a general audience: The Vega Trust Science Videos, “Tick, Tick Pulsating Star: How We Wonder What You Are?” The Cagliari Pulsar Group, “What’s a Pulsar” 
  29. Thibault Damour, “Binary Systems as Test-beds of Gravity Theories,” Physics of Relativistic Objects in Compact Binaries: From Birth to Coalescence Astrophysics and Space Science Library 359 (2009): 1–41. 
  30. When pulsars do not emit radio waves, but more energetic photons in the form of gamma rays, the emission mechanism differs from the simple lighthouse model. Nancy Atkinson, “‘Lighthouse’ Analogy No Longer Works for Pulsars,” Universe Today, January 6, 2009. 
  31. The first experiments to come online will be the Laser Interferometer Gravitational-wave Observatory and the European Gravitational Observatory. For an overview of the scientific possibilities for these experiments see Clifford Will, “The Confrontation between General Relativity and Experiment,” Living Reviews in Relativity 17, no. 4 (2014). 
  32. Paul Steinhardt, along with Alan Guth and Andrei Linde, is widely recognised as one of the fathers of inflation theory. However, in recent years he has become one of the theory’s most prominent critics. Steinhardt’s website includes a page dedicated to criticism of inflationary cosmology
  33. Paul Steinhardt and Neil Turok, “A Cyclic Model of the Universe,” Science 296, no. 5572 (2002): 1,436–39. See also Steinhardt’s web page on cyclic cosmology, which provides both technical and popular resources. Another alternative that seems to fit current data is string gas cosmology: Robert Brandenberger, “String Gas Cosmology after Planck,” doi: arXiv:1505.02381. 
  34. Ross Andersen, “In the Beginning,” Aeon, May 12, 2015. 
  35. The story of BICEP is beautifully told by a BBC documentary for their Horizon series: “Aftershock: The Hunt for Gravitational Waves.”