, , , , , ,


The sun rises at the South Pole, where the BICEP team’s telescope collected data.

Great surprises in science don’t just happen–they’re engineered.

When researchers announced earlier this week that they might have made what is essentially the scientific breakthrough of the year–echoes from the earliest fraction of a second after the Big Bang known as primordial B-mode Polarization –it seemed to come out of left field. Similarly large announcements, like the discovery of the Higgs boson, generally have followed months of speculation, rumors, and even leaks.

It’s standard practice for researchers to keep tight-lipped about their results. No one wants to cavalierly mention half-finished data to a colleague and give them the wrong impression or worse, tip off a rival project. Yet scientists are human, and humans love to gossip. In this world of science blogs and Twitter, the (Background Imaging of Cosmic Extragalactic Polarization) BICEP2 collaboration maintaining secrecy so well is almost unheard of.

The researchers didn’t use some sort of unhackable connection and they didn’t pass notes written in indecipherable code. They had to rely on each other to keep quiet until they could casually drop a major discovery on the world. Here’s how they did it.

The search for primordial B-mode polarizations started in 2001 over a game of tennis. Physicist Jamie Bock, then a researcher at JPL (Jet Propulsion Laboratory | Caltech) had a regular match going with an astrophysics postdoc named Brian Keating (now at the UC San Diego – Department of Physics).

“Brian would bug me about a degree scale polarization experiment,” said Bock. “And after every match I’d go ‘Uh huh, OK, sure.’ But after a while he started to convince me this was worth doing.”

“That’s when we were all like, ‘Wow, crap, maybe it is real.’”

At JPL, Bock had been working on specialized detectors (still under development at the time) that, if placed in a small telescope at the South Pole, could potentially detect the primordial B-modes. He approached the late Caltech physicist Andrew E. Lange with a proposal to search for this signal. Well known in the field, Lange helped Bock assemble the team of scientists, post docs, and grad students to achieve their goal.

“With his help, the whole project just took off,” said Bock, who became one of BICEP’s four principal investigators.

The theory of Inflation, which posits that the universe went through a massive expansion very early in its history, is about 30 years old. Scientists have long known the event, if it had happened, would have left its mark on the cosmos in the form of characteristic twists in light arriving from 380,000 years after the Big Bang known as the Cosmic Microwave Background (CMB). But the hunt for primordial B-modes was known for at least two decades in the field as a “high risk, high reward” experiment.

The history of the universe, showing inflation and its signature on the CMB. Click to enlarge. Image: BICEP team

The history of the universe, showing inflation and its signature on the CMB. Click to enlarge. Image: BICEP team

To detect the signature of inflation, a telescope would need to discriminate minute changes in temperature on the order of 10 millionths of a degree. Some versions of inflation also might have produced a signal that was virtually imperceptible. But if they could be found, these primordial B-modes would open up a whole new world of physics.

Besides providing proof for inflation, the signal would allow scientists to probe unheard of energy levels in the early universe and provide Einstein with another notch in his science belt by proving gravitational waves were real.

“People said, ‘Collect B-modes, collect your Nobel Prize,’” said astronomer Christopher Sheehy, a graduate student at the University of Chicago who joined the team in 2006 under cosmologist Clement Pryke, now at the University of Minnesota. (Full disclosure: Sheehy, as well as another grad student mentioned in this piece later, Neil Turok (who worked on an inflationary model of his own with Stephen Hawking in the 1990s), were undergraduate classmates of mine at the University of California, Berkeley).

The first BICEP project ran from 2006 to 2008 at the South Pole. Though it did not include the specialized detector Bock had been developing at Jet Propulsion Laboratory – NASA (JPL), it was the first step in collecting data and understanding what the team was searching for. A successor experiment incorporating new the new detector technology, BICEP2, started in 2010 and gathered data until 2012.

“We saw hints in these early stages,” said cosmologist John Kovac of Harvard, another principal investigator on BICEP. “But I would say the process for us was of a slow emergence of this signal from the noise.”

Some bloggers were speculating before the team’s announcement that they’d have to be 007-level spies to keep the results under their hat.

Everyone on the team started with a large dose of skepticism about what they were seeing. They didn’t want to get overly excited and unintentionally skew their results. Moreover, they still were not sure at this point that the primordial B-modes could be seen at all.

“We were trying to stay kind of logical and impartial, trying to look at what the data’s telling us,” said Jamie Tolan, a physics graduate student at Stanford University who joined the team under the final principal investigator, physicist Chao-Lin Kuo, in 2007.

BICEP2 cosmic-evolutionCosmic evolution from the Big Bang to today

If worst came to worst, and signal turned out to be nothing, the BICEP team figured it would at least set tighter limits on what other collaborations should one day see. But as more data came in “we realized there was something there,” said Bock.

The team worked hard to ensure this wasn’t some other signal they were detecting erroneously. The telescope and instruments, for instance, can be a source of noise that might happen to mimic the primordial B-mode polarizations.

“We were checking and cross checking, and doing high fidelity simulations,” said Sheehy. “We needed to really understand every effect of our instruments–to understand it down to a level of detail that’s pretty rare.”

The fact that they had used two different detectors, an older technology on BICEP1 and newer one on BICEP2, helped assure them that the instruments were not likely to be a source of problems. One type of instrument could show a particular error, but for two entirely different technologies to do so was unlikely. The collaboration was by this point also running a successor to BICEP2, known as the Keck Array, which provided five times the power of BICEP2. Data from this new telescope helped them check their previous work.

Nearly a year ago, in April 2013, the team got together for a three-day group meeting at Harvard. There, they shared their latest analysis and ideas, trying to stump each other with alternate explanations that could account for their signal. They debated the findings for two days. By coincidence, the final day of the BICEP team’s discussion came on the day of the Boston Marathon, which was marked by a bombing that killed three and injured hundreds of others.


Graduate student Justus Brevik tests the BICEP2 telescope. Image: Steffen Richter, Harvard University

“After that, the whole city was on lockdown,” said Kovac.The team couldn’t get together in person so the PIs got on the phone together. They went around, polling each other on whether or not they thought the signal was real. Among them were optimists and pessimists.

“One person was 80/20, another of us was 50/50,” said Bock. “And whoever didn’t think it was real, we’d ask why they didn’t think so. And then we’d decide on what test we had to do to convince them.”

Bock said that, for him, this meeting was the real watershed moment. “That’s when we were all like, ‘Wow, crap, maybe it is real.’”

The team realized that, no matter how things shook out, they didn’t want to spread false rumors. They went into quiet mode. As they conducted their tests, they started upping their internal security, changing passwords, making new email lists for team members to communicate on. By December, the team had convinced one another. Now they just had to convince the world.

When it comes to keeping a secret, the BICEP collaboration has one major advantage over other physics groups: It’s small. The entire team is around 50 people, and the core analysis group numbers about 20. Unlike the hundreds of researchers working on the Planck space telescope or the thousands of physicists involved in finding the Higgs boson at the LHC, the BICEP members had a fighting chance to keep a lid on their findings. “We could fly under people’s radar,” said Tolan.

BICEP2The BICEP2 telescope at the South Pole

Sheehy recalled there were no specific disciplinary actions that team used to enforce their secrecy. “Everyone was just on board with, ‘Lets not spill the beans.’”

Because of the often leaky nature of science, some bloggers were speculating before the team’s announcement that they would have to be some sort of 007-level spies to keep the results under their hat. The collaboration finds such ideas pretty silly.

“We simply wanted to present our work to our colleagues in full,” said Kovac. “There was no cloak and dagger stuff.”

But about two weeks before the big announcement, the team had to start letting others in on their secret. Kovac personally delivered a draft of the work to theorist Alan Guth, who helped invent inflationary theory three decades ago.

“That’s I think when the rumors started building up a little more,” said Tolan. “I think it was an inescapable thing once we had to start to tell a wider circle of people.”

Excitement began to build. On Mar. 12, Harvard-Smithsonian Center for Astrophysics sent out a mysterious notice to reporters and the public, stating that they would be hosting a conference “to announce a major discovery” five days later. No further details were given. By Friday, several blogs and a story in The Guardian were reporting on rumors of the discovery of gravitational waves from the beginning of time. Over the weekend, the speculation hit a fever pitch and by the time of the Center for Astrophysics (CfA) announcement, the physics world was waiting.

Keck ArrayThough they knew intellectually that their findings were important, many members of the team were somewhat blindsided by the media attention. “The news became a big deal, and when I was getting 10 text messages from cosmologists friends, it hit me on a visceral level,” said Sheehy. “It kind of snowballed, and we realized it’s exactly as big a deal as everyone said it would be.”

Now that the results are public, they been much discussed and dissected. Cosmologists have been debating the findings on Twitter since Monday. Many scientists are impressed but have also been asking for caution, not getting too excited before another team can confirm that the BICEP2’s primordial B-mode signal is real.

Still, for the team members, it will always be an impressive accomplishment.

“It hit me over the weekend that we go to the South Pole, and we build these telescopes with our hands,” said Tolan. “It’s amazing that you set that up to learn about something that happened a trillionth of a second after the Big Bang. The fact that you can do that is mind-blowing.”