Balloon path through uncharted waters.

(Illustration by Anthony Freda)


From dark matter to gravitational waves to a balloon-borne telescope, scientists discuss how they handle setbacks.

Scientific progress follows a winding path, filled with detours and wrong turns—a natural result of exploring the unknown. Science makes headway by challenging itself, identifying mistakes, self-correcting, and persevering. That’s how alchemy becomes chemistry, astrology becomes astronomy, and belief in the four humors leads to medicine.

UChicago scientists have seen their share of scientific wandering. One describes searching for something that no one is sure even exists, and how not finding it is in fact a discovery. Another explains how skepticism—of historical discoveries as well as his own team’s data—leads to more reliable methods, sensitive instruments, and credible results. And one story is a study in resilience in the face of repeated misfortune, and in how catastrophe can give rise to creativity and improvisation.

Science is not a “lockstep march toward progress,” says Edward “Rocky” Kolb, dean of the Physical Sciences Division. He compares the process to Brownian motion, with ideas bouncing around erratically but with a general direction toward deeper understanding and more correct results. “How do we know what the right direction is? We bump into a wall and say, ‘Oops, that’s the wrong way.’”

High hopes

Angela Olinto improvises when her experiment crashes.

On April 25, astrophysicist Angela Olinto let go of her balloon.

Launched from Wanaka, New Zealand, it rose more than 20 miles into the sky—a stadium-sized super pressure helium balloon, carrying a one-ton UV telescope and Olinto’s hopes to discover the secrets of ultra-high-energy cosmic rays. “I find the most energetic particles exciting,” says Olinto, the Albert A. Michelson Distinguished Service Professor of Astronomy and Astrophysics, “because they challenge our theories on how they became so energized.”

The extremely rare charged particles strike Earth at a rate of one particle per square kilometer per century. When they collide with the atmosphere, they produce a cascade of secondary particles, including neutrinos. If astrophysicists can observe those particle showers, they can look backward and search for their origin.

The balloon’s payload, an instrument called the Extreme Universe Space Observatory (EUSO), was designed to measure the UV light produced when nitrogen molecules in the atmosphere are energized by the cascade and then return to ground state. The balloon was scheduled to carry the fluorescence detector for 100 days, testing the equipment but “mostly collecting data,” says Olinto.

Three days into the flight, the balloon sprang a leak. By day 12, it was at the bottom of the South Pacific Ocean. NASA planned for this possibility and sank the balloon, using a remote termination command to prevent a dangerous descent. NASA’s 30-year-old balloon program had conducted an environmental analysis of an open-ocean landing and designed the payload to act as an anchor, pulling the entire balloon quickly to the ocean floor to protect marine life.

Olinto had no say over if or when the balloon should come down. “We are responsible for the payload,” she says. “The balloon and the flight—that’s all under NASA’s control.” Despite her disappointment, Olinto stays positive. “This was not my worst nightmare. That would have been completing the 100-day flight and finding our equipment doesn’t work well.”

The 13-country EUSO collaboration was able to collect some data, in part because after the leak the researchers changed their strategy to optimize what time they had left. “We had to improvise,” says Olinto.

Normally they would collect data on moonless nights, when the particle shower lights are best observed, and download data when the moon is bright. When the leak was confirmed, they downloaded no matter the moon’s state. Luckily their launch window opened during the new moon, and they collected about 60 gigabytes of data.

The balloon’s leak is one of many setbacks the EUSO project has faced. A version of EUSO was originally designed for the International Space Station (ISS) in the early 2000s, but after the 2003 Space Shuttle Columbia disaster, NASA halted space shuttle missions for more than two years pending the investigation. The shuttle program was then phased out in 2011.

In 2012, when the detector was reconfigured for the Japanese Experiment Module of the ISS and became JEM-EUSO, Olinto was invited to lead the US branch of the 13-country collaboration. But several factors, compounded by the 2011 Fukushima meltdown, made the future of that project uncertain. So JEM-EUSO was broken into several projects, one of which was EUSO-SPB, aboard the super pressure balloon, whose launch was then delayed a month by weather concerns.

“I have been in many situations where it looked like the whole effort was about to dissolve into dust,” says Olinto. Yet she finds those situations filled with creative energy, which she funnels into formulating new approaches. “The goals in research are flexible,” she says, “so the alternate path and the final destination are redefined when challenges are overwhelming.”

Olinto’s new plan is to build another telescope and add a neutrino detector. The project’s second generation, EUSO-SPB2, received a NASA award in September. “No one has seen ultra-high-energy neutrinos before,” she says. The second flight will allow EUSO to collect more data and test the neutrino instrument’s capabilities. “It will be easier to predict and prepare for what can go wrong, learning from the first flight, where lots of things went wrong.”

Second time’s the charm. And the fourth. And the fifth.

Daniel Holz, SM’94, PhD’98, explains fake gravitational waves.

On Monday, September 14, 2015, at 4:51 a.m. CDT, the twin Laser Interferometer Gravitational-wave Observatory (LIGO) detectors—in Hanford, Washington, and Livingston, Louisiana—picked up the signal of gravitational waves. Produced by the collision and merging of two massive black holes, it was the first observation of the ripples in space-time that Albert Einstein had predicted a century earlier.

Five months after the detection—once scientists, including UChicago associate professor of physics Daniel Holz, SM’94, PhD’98, had checked, rechecked, and triple-checked the data—they announced their results to the world.

As it turns out, however, this wasn’t the first time LIGO had been through this drill; it was just the first time that it turned out not to be a drill.

Five years earlier, before Holz joined the collaboration, a less sensitive previous incarnation of LIGO had picked up what appeared to be gravitational waves. The collaboration had gone through all the usual steps with the detected event: “It was studied, taken apart, everything, hundreds and hundreds of people involved” over several months, says Holz. A paper was drafted; the decision was made to submit it for publication. “We're talking about people arguing about the title of the paper,” Holz says—it was that close to done.

There was just one problem: there had been no event.

The initial signal had been a “blind injection,” a test designed by a sworn-to-secrecy team within LIGO to see if the equipment—and most important, the scientists interpreting the data—could distinguish between a false positive and an actual event.

“The answer,” Holz says, “was, ‘No, this isn’t real.’ The answer was, ‘We’re not publishing this. We haven’t just detected gravitational waves, and no one’s getting a Nobel Prize.’”

It might seem like “a complete waste of time,” Holz says of the negated months of work, but it’s “actually useful. It makes you go through the whole process” and ask, what went wrong, what did they get right, and how could everything be improved? It keeps the scientists on their toes.

Such tests are standard in the field of gravitational waves research, and an understandable precaution when you’re working to confirm a key part of the general theory of relativity. The abundance of caution is part of the legacy of the first scientist to claim to have detected gravitational waves, Joe Weber of the University of Maryland—the “father of the field,” Holz says, and “an absolutely brilliant experimenter.” In 1969 Weber published a paper in Physical Review Letters that described what he had detected.

But the signal he had found was “at least five orders of magnitude too loud,” Holz explains. Others “could not think of any way from the theory side that there really could be waves that were that loud.” No one else was able to reproduce Weber’s results. Nonetheless, he remained convinced and continued to make more “detections” throughout his life.

Weber’s example “set a particular tone” to the search for gravitational waves, Holz says, and so the goal for LIGO was “to have our detection, especially our first detection, to be so clear, so impressive, that no one could possibly doubt what we’ve done.”

After the false alarm of the blind injection, which came during the era of “initial” LIGO, improvements in the detectors made them far more sensitive. By September 2015 “advanced” LIGO was ready—or almost. In fact, at that point the new equipment was not officially online. “We were still fiddling with the machine,” Holz says. “We were going to turn it on very soon.”

So when the detection came through, everyone assumed it had to be an injection. That’s when they received word from the top: the blind injection system was not yet up and running. And if such a “perfect event” wasn’t an injection, it could be only one thing.

“We still ripped it apart,” says Holz. Without the blind injection system up and running, it was even more important to make sure they weren’t fooling themselves. “It was five months of a thousand people doing their very best to figure out how this might not be real.” But it was real. “We couldn’t make the event go away.”

More gravitational waves have followed—confirmed detections in December 2015 and January 2017. Conservatism, however, still rules: an October 2015 detection is classified only as a “candidate” gravitational wave because it wasn’t loud enough for the collaboration to be confident.

To this day, however, LIGO has yet to switch on its blind injection system. “Because we've seen real events, we know it’s working,” Holz says. So the last thing they need is fake signals to analyze. “At this point it’s becoming difficult to keep up with the real events that keep showing up.”

Process of elimination

Rocky Kolb searches for the mysterious particle.

Astrophysicists theorize that about 85 percent of the universe’s mass is dark matter, which can be detected only through its gravitational effects. Galaxies and galaxy clusters spin so quickly that they should have torn themselves apart based on their observable matter. Something is holding them together, but no one knows what.

Scientists know much about what dark matter is not: It is not the visible stuff of stars and planets. It is not dark clouds of baryonic (ordinary atomic) matter, which can be observed absorbing radiation passing through them. And it’s not antimatter, which would produce gamma rays when it annihilates with matter. So what is it?

One hypothetical candidate is WIMPs—weakly interacting massive particles that don’t interact much with ordinary matter, proposed more than 30 years ago. As a graduate student at the University of Texas, Austin, in the 1970s, Kolb, now the Arthur Holly Compton Distinguished Service Professor of Astronomy and Astrophysics at UChicago, helped lay the foundations for WIMPs by exploring the limits to weak interaction.

WIMPs may bepart of the concept of supersymmetry, which fills gaps in astrophysicists’ understanding of known particles and forces. The idea says that each fundamental particle has an as-yet-undiscovered superpartner. When scientists use the properties of the lightest supersymmetric particles—WIMPs—and calculate how many would still exist after the big bang, that number matches the amount of dark matter seen (or inferred) today.

LUX dark matter detector
At Sanford Underground Research Facility, a mile beneath the Black Hills of South Dakota in a former gold mine, the Large Underground Xenon dark matter detector continues to search for WIMPs.

But so far no detectors or colliders have been able to shed light on WIMPs. So does Kolb still think they’re the answer? “I think we’ll be surprised, that the answer will come out of left field,” he says.

What’s advantageous about the WIMP hypothesis says Kolb, is that it’s falsifiable. British philosopher Karl Popper’s concept of falsifiability states that theories are scientific only if it is possible, in principle, to prove them false, and that empirical science is never confirmed, only incrementally corroborated through absence of disconfirming evidence.

Another dark matter candidate—ordinary matter in the form of black holes, neutron stars, or brown dwarfs called MAssive Compact Halo Objects, or MACHOs—was falsified in 2004 through the discovery of a galaxy cluster that doesn’t behave in accordance with the hypothesis.

“Maybe we’re on the verge of falsifying WIMPs,” says Kolb, which would be a form of discovery.

He cites the famous failed experiment of Albert Michelson, founder of UChicago’s physics department, and Edward Morley to establish the existence of “ether,” the medium they believed filled space and was required to transmit light. In the process of failing, they established the speed of light as a fundamental constant, and their work eventually led to the theory of relativity.

So discovering that WIMPs aren’t the explanation for dark matter would point astrophysicists in other directions. But scientists “should completely exhaust the possibilities,” Kolb says, before making that call.