Five years after BICEP2

In March 2014, members of the BICEP2 collaboration held a press conference to announce a finding worthy of a Nobel Prize. Using a sensitive microwave telescope near the South Pole, the researchers had detected a polarization pattern in the cosmic microwave background (CMB) that they attributed to gravitational fluctuations during the first 10^–35 seconds of the universe’s existence. As Physics Today reported soon after , the discovery seemingly marked “both a triumph of inflationary theory and the first real evidence of the quantum nature of the gravitational field.”

The result didn’t hold up for long. By that fall the Planck collaboration had released data that strongly suggested that galactic dust , not primordial gravitational waves, had imparted the observed polarization signature. When the BICEP2 team published a joint analysis with Planck in early 2015, it could say only that there might be a subtle primordial signal hidden deep within the dusty noise .

Five years after the flashy press briefing, it appears that nobody will be making a similar announcement anytime soon. Steadily improving measurements from the BICEP team, which is still leading the hunt, and other collaborations reveal that if the inflation-era imprint on the CMB exists, it is far subtler than it seemed in March 2014. Detecting it likely will require a more advanced set of instruments to achieve the requisite sensitivity—or to show that we have to rethink how we view inflation.

Though the BICEP scientists received plenty of criticism for how they disseminated their initial findings, few doubted the reliability of their raw measurements. The BICEP2 telescope, fitted with 512 cryogenic transition-edge sensors , precisely mapped the CMB emanating from a small slice of sky at a frequency of 150 GHz. The telescope detected curl-like B-mode polarization, which can be produced by a tensor field, such as the gravitational field during the era of inflation. But that’s not the only way to generate B modes. Dust can imprint a similar pattern, and in 2014 there were no reliable maps or theories that fully captured dust’s polarization signature at various frequencies. There still are none.

Even as the robustness of its reported signature steadily deteriorated throughout 2014, the BICEP team continued to improve its instruments. The researchers installed new detectors at 95 GHz and 220 GHz in attempts to boost sensitivity and isolate the polarization due to dust, whose influence depends on frequency. BICEP2 has since been replaced by BICEP3, which is scanning the southern skies at 95 GHz with 2560 sensors. The collaboration also operates the 220 GHz Keck Array, essentially a network of five BICEP2 telescopes.

Last year, by combining data through 2015 from BICEP/Keck with those from Planck and the long-retired Wilkinson Microwave Anisotropy Probe satellite, researchers set an upper limit of 0.06 for the tensor-to-scalar power ratio r, which can be used to calculate an energy scale for inflation. That’s a far cry from the team’s 2014 result, which found that r most likely falls between 0.15 and 0.27 and is inconsistent with 0.

As the maximum possible value of r creeps downward, the window gets smaller for some theories of inflation. Models such as natural and quadratic inflation that were popular several years ago no longer seem tenable, says theorist Marc Kamionkowski of Johns Hopkins University. Plenty of workable theories remain, but that will change if r turns out be 0.01 or lower. “It’s pretty hard to get a value less than that in any basic textbook model of inflation,” Kamionkowski says. And if r is shown to be less than 0.001, “what we teach would no longer be viable. It wouldn’t mean that inflation didn’t happen, but it would mean that it didn’t happen the way we now think it did.”

The recent improvement narrowing down r is a testament to the skill of the experimentalists, but it also complicates their job going forward. Determining ever-lower limits is useful to the cosmology community, but the ultimate goal of BICEP/Keck and its competitors is to detect the imprint of cosmic inflation, if it is indeed detectable. “The bar for satisfying ourselves and others that we have a genuine detection gets harder to reach as r gets lower,” says the University of Minnesota’s Clement Pryke , a coleader of the BICEP/Keck team. “We want an experiment that could potentially make a detection.”

To do that involves not only increasing the sensitivity and resolution of detectors but also boosting efforts to remove the effects of foreground sources. By next spring the BICEP/Keck team hopes to begin operating the BICEP Array, which when completed will use four BICEP3 receivers to observe at six different frequencies and help sort out the influence of dust. The collaboration is also teaming up with a competitor, the South Pole Telescope (SPT), to quantify the polarization caused by gravitational lensing. Due to its higher angular resolution, the SPT is better than the BICEP and Keck instruments at analyzing the effects of galaxies and other massive structures, which twist less-interesting E-mode polarization into B modes.

The joining of forces will likely become a theme as the hunt for B modes gets more complicated. Backed by $40 million in funding from the private Simons and Heising-Simons Foundations, the Simons Observatory will use a combination of lens-based telescopes and radio dishes to probe the CMB in Chile’s Atacama Desert. The collaboration includes researchers from the current Atacama Cosmology Telescope and POLARBEAR projects, which are also located in Chile, and is co-led by Brian Keating, who wrote a book about his whirlwind experience as a member of BICEP2.

Meanwhile, researchers from nearly all the major B-mode projects are planning CMB-S4 , a next-generation observatory with more than half a million detectors. The telescopes in the array would complement one another by focusing both on small slices of sky, like BICEP/Keck does, and on larger swaths, which reduces the influence of gravitational lensing.

Although competition still exists between various groups, the community seems to grasp—especially after the events five years ago—that no single experiment is going to provide smoking-gun evidence on its own. Says Kamionkowski: “We’ll need measurements from multiple experiments to get a result that we really believe in our bones.”