The idea was too beautiful to be wrong.

That you could start with nothing, apply some basic laws of physics, and get a universe out of it—a universe that was uniform on the largest scales but replete with the lumps and bumps we call stars and galaxies, a universe, that is, that looks like ours —well, it didn’t matter that the theory didn’t quite work at first. It was just too beautiful to be wrong.

Inflation. In Alan Guth’s original version of the theory in 1980, the nothingness at the beginning of time wasn’t really nothing at all. It was a field, the inflaton, and it teetered at the edge of a cliff, momentarily stable but not in its most stable, or lowest energy, state. This gave spacetime a negative pressure, creating a kind of anti-gravitational force that would push outward, sending the inflaton—that nascent field that would give birth to inflation—plummeting toward stability, causing the universe to expand exponentially, growing a million trillion trillion times bigger in the blink of an eye.

It was creation nearly ex nihilo —all you needed was the tiniest speck of a universe and inflation would transform it into something truly cosmic. There was just one problem: the plunge to the lowest energy state was a kind of phase transition, like water vapor condensing to liquid, and the transition would dissolve the inflaton into a sea of bubbles—pockets of lowest-energy regions—which would eventually collide and merge, collisions that would leave astronomical upheavals more disfiguring than anything we see on the sky today.

Then, in 1981, Andrei Linde saved inflation from itself. He suggested that we didn’t have to worry about those bubbles because inflation could make them so big that our entire universe could fit inside just one of them. It didn’t matter what happened out at the edges or beyond—we’d never see it anyway.

There was just one problem. The smooth, scarless space inside the bubble was too smooth, the density of matter so perfectly uniform that nothing so lumpy as stars or galaxies could ever form. It was Linde’s friend and fellow physicist Slava Mukhanov who had the solution: quantum fluctuations.

Inflation was creation nearly ex nihilo.

Quantum fluctuations are born of Heisenberg’s uncertainty principle, which says that certain pairs of physical characteristics—position and momentum, time and energy—are bound together by a fundamental elusiveness, wherein the more accurately we can specify one, the more wildly the value of the other fluctuates. The universe cannot be perfectly uniform—uncertainty will not allow it. At a precise moment in time, energy varies recklessly; at a well-defined position, momentum soars and swerves. Precise moments and well-defined positions normally mean tiny scales of time and space, but inflation blows all that up. Inflation, Mukhanov told Linde, could take these tiny quantum fluctuations on the order of 10 -33 cm and stretch them to astronomical proportions, creating slight peaks and valleys throughout space and laying a gravitational blueprint for what would eventually become a network of stars and galaxies.

Still, Linde wasn’t satisfied. Getting inflation to start and end in just the right way required the whole thing to be improbably fine-turned. It was beautiful, but unnatural. There would be two more years of work before he found the solution: chaos. Inflation didn’t require fine-tuning, he realized; it didn’t need to teeter on a cliff’s edge. If the inflaton started off in a highly-probable and totally random state, then somewhere amongst the mess, there was bound to be a region with the right properties to spark inflation. From a sea of chaos, a vast island of order would emerge.

That’s where the universe stood in the cold Moscow winter of 1986. Gorbachev had recently taken office as the General Secretary of the Communist Party and had just set into motion the perestroika —the restructuring of the Russian political, economic, and educational systems. For physicists like Linde, this engendered a strange silence. The old system for getting academic papers published abroad had been scrapped, but it hadn’t yet been replaced by a new one. So while inflation was being developed in the U.S., Russian physicists were forced to wait.

Linde waited in bed. The doctors told him he was perfectly healthy, but he felt awful nonetheless. He was passing the time reading detective stories when the phone rang. It was the administration from the Lebedev Physical Institute, where he worked. They told him he was to travel to Italy to give a public lecture. He didn’t want to go. Under Gorbachev, Linde was allowed only one trip abroad each year, and he wasn’t about to waste it on a public lecture where he wouldn’t be working with other physicists or learning anything new. He told them he was too ill to travel. You are ill today, they said, but you’ll likely be healthy again soon, no? Or are you saying you are unable to go abroad at all?

Linde grew scared. He knew if he said that he was unable to go abroad, they might never let him leave again— ever . He needed to prove that he could make the trip, and quickly. It was a Friday. He needed to get to the Hospital of the Academy of Sciences in order to obtain a certificate of health, but he was just learning how to drive and couldn’t risk a battle with the Moscow ice. He decided to pay for a taxi, a financial decision that didn’t come easy. Over the weekend he prepared the necessary travel documentation, and on Monday invested in another taxi ride to the Institute. He paid secretaries to immediately type up his paperwork, which he then ran to every corner of the Institute to get every last signature required. That bureaucratic nightmare ought to have taken a month and a half, and he accomplished it in four days. He dropped off the papers, went home, and collapsed into bed. He didn’t get up for two days.

Soon the phone was ringing again. The trip was set, they told him, but the Italians wanted to see his lecture ahead of time—the day after tomorrow. Suddenly, Linde realized he had a golden opportunity. He could get around the systemless system and publish abroad! Instead of handing over his public lecture, he could write a new paper, give it to the powers that be and they would send it abroad for him—by diplomatic mail, no less. There was just one catch: he had half an hour to do it. It was the only way to get it typed up in time.

Linde sat with his head in his hands, rolling it from side to side. Think, think . He felt like a compressed spring—he would either bounce to new heights or break under the stress. He knew that theorists can’t simply order up good ideas at will—physics doesn’t work that way. But today, he thought, it was going to have to.

Thirty minutes later, he had come up with the theory of the chaotic self-reproducing inflationary multiverse. It was his greatest piece of work.

Linde’s new theory reached beyond the bounds of the bubble. In his earlier version, our little patch of inflationary universe would arise from some small stretch of chaos. But while our universe was growing, what was happening behind the scenes? Surely there would be other regions where inflation could crop up. They’d be rare, but it didn’t matter—they would grow so big so rapidly that they would soon dominate the landscape. Each inflationary region creates more of itself—it’s self-reproducing. The process ends locally within each island universe, but on the largest scales it carries on, producing universe after universe after universe. In a half hour, Linde had taken our single universe, once the whole of everything there ever was or would be, and duplicated it, multiplied it, mutated it, sent it through a sequence of funhouse mirrors until it emerged on the other side a mere speck again, a humble, lone bubble in an infinite and growing multiverse.

Seeing gravity waves…it would be like a fish seeing water.

When he first developed the idea of inflation, Linde never for a second thought that it would be technologically feasible to test it. In principle, there were ways—you could look for the tiniest temperature fluctuations in the remnant heat from the Big Bang, those tiny quantum fluctuations that seeded the stars and galaxies, but that was a precision measurement he could barely fathom at the time. And if you wanted to dream even bigger, well, there ought to be something even more fundamental—quantum fluctuations of spacetime itself, primordial gravity waves. Seeing gravity waves…it would be like a fish seeing water. And seeing primordial gravity waves…well, it’s not just any water, it’s the first water, the origin of water, the origin of everything. But the technological skill that it would take to make that kind of measurement—it was downright unthinkable.

On good days, he didn’t care. He knew the theory was right, he knew it in his bones. He knew it with the same kind of certainty that Einstein had about general relativity: When observations of the 1919 eclipse came in, proving that gravity bends light just as general relativity predicted, a reporter asked Einstein how he would’ve felt had the experiment turned out differently. “I would have felt sorry for the dear Lord,” Einstein replied, “because the theory is correct.”

The Device

There was a problem with the antennas.

When Chao-Lin Kuo arrived at NASA’s Jet Propulsion Laboratory in Cañada Flintridge, California in 2003, the BICEP team was trying to implement Jamie Bock’s vision for a new polarization detector in their search for primordial gravity waves. Not that the old detectors didn’t work, but the things were unwieldy. Three copper feed horns, a handmade filter, and two detectors per pixel, all hand assembled. It’s not that they weren’t sensitive—they were nearly as sensitive as you can get. Rather, if they wanted better measurements, they didn’t need more sensitive detectors, they needed more detectors—quickly and cheaply. Bock’s vision was to digitize the whole assemblage and print them on circuit boards with microlithography, creating a kind of mass-producible polarimeter-on-a-chip. If it worked, it would change everything. It would be like upgrading from vacuum tubes to integrated circuits. But the team was stuck. They had designed a beautiful antenna array, but its readings kept coming out wrong.

BICEP2 Detector — A single polarization detector

The plan was to mount the detector to a radio telescope at the South Pole, where it would catch light that’s been traveling through an expanding cosmos for the last 13.8 billion years and measure its polarization, or the direction in which the photons are waving relative to the direction of their motion. If they could pin down each photon’s polarization with enough precision and map them across the sky, they’d have some hope of discerning a pattern known as a B-mode, the signature of primordial gravity waves. Kuo, a 30-year-old postdoc, set to work, putting the array through a host of tests until he figured out the problem: it was because the feed lines were crossed. The array looked like a series of X’s, but at the center of each X, the antennas were picking up each other’s signals and screwing up the reading. He set to work on a new design.

Kuo knew he had to keep the antennas at right angles from one another so they could subtract the horizontal polarization from the vertical and take the difference. And he had to keep them as symmetric as possible, because the difference they were looking for was one part in 30 million. One part in 30 million . All to find a B-mode. How exactly do you make something like that?

When he really thought about it, this thing they were trying to do, this thing they were trying to measure, it pushed the bounds of sanity. But Kuo already had a taste for pulling something like this off. As a grad student back at Berkeley, he had worked on the ACBAR experiment, which took measurements of the cosmic microwave background temperature fluctuations. The idea that you could build something with your own two hands, point it at the sky, and see the faintest details of the nascent universe some 14 billion years in the past…well, you had to see it to believe it. You see the pattern. It’s not an image in a textbook or an idea in your mind— it’s on the sky . You look at it and suddenly you realize that you are one of a handful of human beings who has ever cast his eyes on the Big Bang. Well, it’s not exactly the Big Bang; it’s 380,000 years later–a mere eyeblink in the cosmic course of things, but still. To see back to the very beginning, the very first fraction of the very first second, you need something better than light. You need gravity.

…suddenly you realize that you are one of a handful of human beings who has ever cast his eyes on the Big Bang.

Kuo tried design after design. On some level, the antennae weren’t all that different from the kind you’d find in a cell phone, except this cell phone needed to answer calls from the beginning of time. The antenna array would shuffle the incoming photons down to the focal plane, where electromagnetism would be converted to heat and measured by an ultrasensitive thermometer. If you want to capture a signal that’s been steadily weakened over 14 billion years, you better make sure there’s virtually no heat and zero polarization coming from the instrument itself or it will totally swamp the measurement. That means keeping the detectors cooled to 0.25 Kelvin, just the slightest shiver above absolute zero. In the old way of thinking, the signal had to be transported off the focal plane and out of the cooling element to be read out by some room temperature electronics, but the transmission itself through heat conducting wires could warm the focal plane enough to drown out the signal. So Bock’s idea was to have the signals read by superconducting electronics on the focal plane itself using quantum-scale magnetic sensors developed at the National Institute of Standards and Technology in Colorado.

In the meantime, they deployed the old detector to the South Pole in an experiment they named BICEP1. For three years, from 2006 to 2008, it would collect that nascent light and look for the slightest patterns of polarization.

BICEP2 Focal Plane — The BICEP2 focal plane

Back at JPL, it was Kuo’s fifth design that stuck. He built an antenna array that looked like a series of H’s, with spaces between the vertical and horizontal lines to avoid having the feed lines intersect. Once the array had been fabricated at JPL, it was time to put them to the test. Kuo placed them carefully in the cryostat. Then he waited.

It would take several days to get things cold enough. First, liquid nitrogen would cool it down to 77 Kelvin. Then the liquid helium would kick in, lowering the temperature to 4 Kelvin. Finally, a few cubic centimeters of helium-3, a rare isotope. With helium-3, you have to tread carefully. The stuff is expensive; as a byproduct of plutonium production, it’s a controlled substance.

While Kuo waited, he thought about inflation. If that exponential expansion really gave birth to the universe, it ought to have taken quantum fluctuations in spacetime and blown them up across the sky. Some 380,000 years later, the photons that make up the cosmic microwave background radiation would have navigated that same warped spacetime, a journey that would imprint itself uniquely in their polarization. Find the B-mode polarization and you’ve found inflation’s smoking gun. Looking around the lab, he wondered if he was the only one worrying about inflation. These guys were hardware wizards—they want to build cool things. Most of them didn’t have a lot of faith in theory. Kuo respected that. But for him, he needed to understand why he was doing this. Yes, he wanted to build a kickass detector. But he also wanted to know how the universe began.

Once the helium-3 had everything cooled to 0.25 Kelvin, Kuo had to test the things, to see if they worked and to diagnose any problems. Start by sticking something room temperature in front of it and see what temperature reads out. Then something cooled with liquid nitrogen. Shine a source of microwaves at it, rotate their polarization, watch what happens. He ran every calibration test he could think of. The antenna array worked.

Kuo had transformed Bock’s vision into a groundbreaking—and more important, functioning—detector. Because they used lithography, they could pack 512 of them on the focal plane, which meant BICEP2 would achieve the same sensitivity as BICEP1 in one-tenth the detection time, much like a bigger camera sensor can capture more stars at night. Kuo’s timing couldn’t have been better. BICEP1 was going off-line and the new technology had to ship out on the first flights of the year to Antarctica in September.

Despite the pioneering technology, the truth was, no one on the team seemed to believe that a detection was in the cards. Even if inflation were correct, there was a good chance that primordial gravity waves would be way too small to measure. They just thought they’d use the telescope as a proving ground for the technology so that later it could be confidently incorporated into a next-generation space satellite. Satellites are expensive, and if something breaks once it’s up in orbit, you’re out of luck. So the BICEP2 team figured they’d take the technology out for a terrestrial test drive; in the meantime, they could place more upper limits on the amplitude of gravity waves and constrain some inflationary models in the process.

No one on the team seemed to believe that a detection was in the cards.

The physicist Andrew Lange had said that this was a wild goose chase. Still, Kuo couldn’t help hoping. Every once in awhile, he figured, you catch a goose. When Penzias and Wilson first discovered the cosmic microwave background in 1964, they thought it was literally pigeon shit. At least the BICEP2 team knew what they were looking for.

In the middle of all that, Kuo had moved up the coast, from Pasadena to Palo Alto. He took a position at Stanford University, where he recruited an eager young grad student named Jamie Tolan to work with him on the measurement. One day, Tolan approached his advisor—he was writing a proposal for a NASA graduate student fellowship, and he asked Kuo to read the draft. In the proposal, Tolan laid out the goal of BICEP2: to see just how elusive primordial gravity waves are. Kuo smiled at Tolan. That’s not it , he told him. The goal is to detect them .

The Questions

Linde had wanted to be a geologist. His father was a radio physicist, his mother an experimental physicist who studied cosmic rays. The younger Linde wanted to do something different, something tangible. Something like rocks. But during the summer vacation between 7 th and 8 th grade, the Linde family drove from Moscow to the Black Sea. For a week, Linde sat in the back seat reading. He had brought two books: one on stars and the universe, the other on Einstein’s theory of special relativity. When they arrived at the Black Sea, three physicists stepped out of that car.

At Moscow State University, Linde sought his colleagues’ advice: should he be a theorist or an experimentalist? The truth was, he didn’t think he was that great at calculation. He did, however, possess a certain intuition coupled with an obsessive mind. Once he became interested in a question, he couldn’t stop thinking about it. Linde soon realized he wasn’t nearly as impressed by measurements as he was by explanatory power. He didn’t want data—he wanted answers . Answers to big questions, the biggest: What happened when he was born? What will happen when he dies? What is it to feel, to think, to live, to exist? But he figured he’d start with simpler questions, the kind with more straightforward answers, like, how does an airplane fly? He promised himself he’d get to the hard ones eventually. There was no denying it. He was a theorist through and through.

Eventually the hard questions snuck back in. When Linde came up with chaotic eternal inflation in that fateful half hour, he immediately realized the implications. In an infinite multiverse where physical constants can vary from one universe to the next, everything that can happen will happen—an infinite number of times. Every possible world, every incarnation of reality, every possible version of you living every possible version of your life. What then does it mean to want something, to do something, to be something? It was a vertiginous thought, but Linde didn’t let it get to him. So what if there were infinite Andrei Lindes? If I killed myself, he figured, it’s not like I’d survive as a copy—my death would simply become the moment that I was no longer identical to my copy, because I, unlike him, would be dead.

In any case, it wasn’t clear that the copies existed in any meaningful way. That was the thing about quantum mechanics—the very nature of things seem to be determined by what an observer can measure. In the world of classical physics, you could have two baseballs that were identical in every way, and yet it’s fair to say that there are two of them. In the quantum realm, if you have two indistinguishable particles, you only have one particle. Wheeler and Feynman had emphasized that—in a sense, they said, there’s only one electron in the universe. Linde could never quite shake that.

Even those quantum fluctuations—the very fluctuations that gave rise to the stars, polarized the microwave light, and created universe after universe—they are determined directly by what an observer can measure. Position and momentum, time and energy—these partners bound by uncertainty are so bound because the accurate measurement of one precludes the accurate measurement of the other. A particle doesn’t have a simultaneous position and momentum because an observer can’t measure a simultaneous position and momentum. Gravity waves are waves of uncertainty—uncertainty not only of existence but of observation. It was a fact that seemed to suggest that observers play some deep role in the nature of reality, a fact that Linde kept tucked away in the back of his mind. What is it to feel, to think, to live, to exist? If there was no observer who could simultaneously observe more than one Andrei Linde, then on some level you might say there’s still only one.

Despite this, Linde was convinced that the existence of all those parallel universes held great explanatory power. While the multiverse was ultimately governed by the same laws of physics—by quantum mechanics and relativity, by inflation itself—each universe would be born with its own local sub-laws, a set of accidents that would determine its geometry, its physical constants, its particles, its forces, its own unique history. Inflation meant diversity. And diversity, Linde realized, was its own kind of explanation.

So many features of our universe appear inexplicably fine-tuned for the existence of biological life. Change the strength of a force here or the mass of a particle there and poof! —no stars, no carbon, no life. Such coincidences demand explanation, and inflation had one: the strengths and masses vary from universe to universe, and we just happen to find ourselves in the one in which we can live. The inflationary multiverse may not have been predictive or observable, but it was explanatory . It could explain the illusion of design, the comprehensibility of the cosmos, the unreasonable effectiveness of mathematics. It could explain why the cosmological constant is so small and why the universe is so big. It could explain why we are here, why anything is here, because at the end of the day, Linde knew, physics isn’t really about the universe.

Linde didn’t like being told what to think.

The mass of the electron is 2,000 times lighter than that of the proton. Why? Well, if it were ten-times heavier or ten times lighter we wouldn't be here to ask. Spacetime has four large dimensions. Why? Well, any more dimensions and the gravitational force between two objects would fall off faster than r -2; any fewer and general relativity couldn't support any such forces all. Either way, you've got no stable planetary systems and no life.

Such explanations are called “anthropic,” and they made people nervous, the theoretical physics equivalent of “just because.” Colleagues told Linde he shouldn’t think about such things, but he didn't like being told what to think. When he decided to include a section on the anthropic principle in the cosmology book he wrote, his editor in Moscow told him to take it out. If you leave it in, she said, you’ll lose the respect of your colleagues. Yes, Linde replied, but if I take it out, I’ll lose my respect for myself.

As far as he was concerned, the metaphysical is always brought into the fold of physics in the end, and inflation meant that the burden of proof was on those who wished to believe in a single universe. Einstein had once said, “What really interests me is whether God had any choice in the creation of the world.” He wanted the universe to be a singular specimen of logical perfection and uniqueness. Not Linde. Linde wanted diversity, choice. In Russia, they only had one choice of cheese.

At the Bottom of the World

It was Kuo’s fourth visit here, at the bottom of the world, but he still wasn't used to the whiteness of it all. Everything, everywhere—just white. A blank spot on the world, like someone forgot to fill it in. An endless white that makes you think about infinity. He must've been ten years old the first time he thought about it, whether the universe was infinite or finite. That was back in Taiwan—some 8,000 miles from here—where the sun still sets on a summer’s night. It hadn't made sense to him, as a boy, that reality would just come to an end, that there was a place beyond which there is no more place. What if you sat there at the edge and threw a ball? Where would it go? Someone else had made the same argument, he remembered. A philosopher? Now, as a physicist, he knew it wasn't so simple— that the universe could be curved and closed, like the surface of a sphere, finite but without an edge. He supposed he had always been a physicist. Funny how all this white makes you think of that. Of all the colors, he missed green the most. Green and the smell of humidity. He had never realized what humidity smelled like until it was gone.

South Pole takeoff — An LC-130 takes off from the Amundsen-Scott South Pole Station.

It was hard to say how many days he had been here—hard to differentiate time when the scenery never changes, the weather never shifts, and the sun never goes down. Getting here had been an adventure, as usual. He had flown some 15 hours from California to Christchurch, New Zealand, for a stopover at the International Antarctic Center, where he traded his belongings for extreme cold weather gear before boarding an Air Force aircraft and flying another 14 hours to McMurdo Station here in Antarctica. From McMurdo it was another three-hour flight to the Amundsen-Scott South Pole Station on a plane that landed on skis. Stepping out onto the ice sheet, he had marveled again at the sky, so perfectly blue—the clearest sky on the planet.

That’s why they were here. Antarctica is the largest desert on Earth. The altitude gets you up above most of the problematic parts of the atmosphere and the biting cold takes care of the rest—any stray water vapor in the air is frozen out of the sky, leaving microwave light from the early universe to stream through unimpeded. It also helps that the sun only rises and sets once a year.

It was December now; he would be here until Valentine's Day. The sun would set in March. He didn't know how the “winter overs”—the people who stayed here past March—did it, not when -20°F was a warm summer day. Of course, the science station had grown more comfortable lately. It had a sauna now and a greenhouse for growing hydroponic fruits and vegetables. Earlier, they used to give you this weird yellow powder, and you'd mix it with water, fry it up, and call it a meal. Now, you could enjoy fresh produce in the cafeteria then go play on the basketball court or relax in the library or game room.

Between the porthole windows in the doors and the firemen’s lockers lining the corridors, the place looked like the perfect combination of a research ship and a high school. Ship was more accurate—the Amundsen-Scott station, perched on Antarctica's high plateau, stands on stilts to avoid the snow that never thaws atop a glacier some 9,000 feet thick that ever so slowly drifts.

To get to work, Kuo would walk along the ice sheet, across the airplane runway, upwind to the Dark Sector lab, so-named because all white light and radio transmission is forbidden there. The lab was hardly a mile away, but cold, wind, and altitude have a funny way of stretching distance. By the time he reached the telescope, he was queasy and out of breath.

BICEP2 was a refracting telescope with a small aperture—just 26 centimeters. It could afford to be small because the features it was looking for were the size of the full moon on the sky. All of its moving parts were kept inside where it's warm. Only its head poked out through a hole they had cut in the roof. The telescope was focused on a 20° patch of the so-called Southern Hole, the cleanest stretch of sky available with a clear view straight out of our Milky Way. At the South Pole, the same patch of sky just keeps spinning in circles above you; it never slips behind the horizon or disappears from sight. The telescope can stare it down for years and never blink.

BICEP2 observed only photons with a frequency of 150 GHz, filtering everything else out. They had opted for a single frequency because it was the only way to optimize every part of the instrument. When you're trying to avoid dust, which can polarize your light and mimic the signal of gravity waves, 150 GHz is the sweet spot. It’s where you’re most likely to see the clearest signal of gravity waves. The two possible impostors, magnetized radiation from extreme astronomical phenomenon and interstellar dust, rise at low and high frequencies respectively. But 150 GHz is right in the Goldilocks middle. It also happens to be the peak frequency of the cosmic microwave background, the photons that flew out of the dense early universe 380,000 years ago.

The telescope had two lenses that focused the light, a design similar to Galileo's, except that this one fed the light into the most sensitive superconducting detectors ever built. Kuo and his team were here to assemble the thing and then take some calibrations, but even turning a screw was proving to be difficult in the cold.

Once the telescope was up and running it would start collecting data, which it would store temporarily on the computers at the South Pole. But soon a low Earth orbit communications satellite would appear above the horizon and relay the data from the South Pole station to NASA's White Sands complex in New Mexico. From there it would bounce around the U.S. until it landed in a cluster of computers at Harvard University, which the BICEP2 team could later access from California.

California. Kuo wondered what his wife and children were doing back home in Stanford. They were probably enjoying the green, green grass and the warmth of a more fleeting sun.

The Observer and the Observed

California. Linde moved here in 1990 with his wife, Renata Kallosh, and their two sons. A year earlier they had left Moscow for Switzerland, intending to spend a year at CERN before heading back to the Soviet Union. But offers came in while they were there, including a double offer from Stanford University for both Linde and Kallosh, who is a string theorist, and so they changed course and immigrated to the U.S.

In the two decades that followed, evidence for inflation mounted, and, in 2003, cosmologists hit the jackpot. NASA’s Wilkinson Microwave Anisotropy Probe (WMAP)—an 1,800-pound spacecraft that orbited the sun nearly a million miles out— had produced an unprecedented map of the microwave sky, measuring temperature differences in the near-uniform radiation down to one part in 100,000. Those slight hot and cold spots traced quantum fluctuations in the density of matter 380,000 years after the Big Bang, when the microwave light was first emitted. The pattern in the map bore out several key predictions of inflation with astounding precision. Even the inflation doubters were coming around. Now there was just one piece of evidence missing: B-mode polarization, the mark of primordial gravity waves.

Linde wasn’t worried about B-modes. Most versions of inflation predicted them at amplitudes way too small to measure, which meant that even a non-detection could be a strange kind of confirmation, at least for those who already believed. As far as he was concerned, the experimental evidence was already overwhelming. Still, he supposed, on the off-chance they did discover B-modes—well, it would just drive home the fact that quantum mechanics needs to be taken seriously, even at cosmic scales. The beauty of inflation was that it provided the missing link between the tiny quantum world and the largest scales of the universe. We are the great-great-great-grandchildren of quantum fluctuations, he liked to say.

When you try to apply the laws of quantum mechanics to the universe as a whole, you hit a paradox: all things quantum are defined in terms of what an observer can measure, but no one can measure the universe as a whole because, by definition, you can’t be outside the universe. The issue was captured most perfectly in the famous Wheeler-DeWitt equation, which showed that the quantum state of the universe could not evolve in time, stuck, as it were, in a frozen, eternal moment. As Linde often put it, without observers, the universe is dead.

Linde knew that the only way to get time flowing was to observe the universe from here, on the inside. When we look out at the cosmos through a telescope, he thought, we don’t see ourselves in the picture. And so we split the world in two: observer and observed. We make a measurement, and the universe comes to life. It sounded awfully solipsistic, but there it was.

As Linde often put it, without observers, the universe is dead.

Everyone assumed you can talk about “observers” without talking about consciousness—things like Geiger counters or space telescopes—but Linde wasn’t so sure. If you remove subjective experience from the picture, he thought, there’s no more picture. He couldn’t help wondering whether consciousness was the missing ingredient that would make the ultimate theory of physics consistent. The idea was inspired by gravity waves.

Back in the day, physicists thought of space and time as tools that we use to describe the motion of matter—not as things in their own right. It was Einstein who realized that even if you emptied the universe of matter, spacetime itself would remain and could exhibit a behavior all of its own: it could wave. Gravity waves meant that spacetime was equally as real and fundamental as matter itself. Later theoretical developments—namely supergravity—extended the symmetries of this space-time so that matter turned out to be nothing deeper than excitations of the geometry of superspace. In other words, it was spacetime that was fundamental and matter was derived, a tool for describing the excitations of spacetime.

Linde wondered if consciousness awaited the same vindication. Today we think of it as a tool we use to describe the external world, and not as an entity on its own. But what if the external world were empty? What if consciousness was fundamental and the universe derived? Could space, time, and matter together be nothing more than excitations, the gravity waves of consciousness?

What is it to feel, to think, to live, to exist? It was still the only question he really cared to answer. The rest was just details.

The Signal

They must have made a mistake.

They had screwed up the analysis or there was some design flaw they hadn't accounted for yet. A signal this bright—it had to be coming from the instrument itself. There was no way this thing was coming from the sky.

BICEP2 had collected data for three years and now the team had set out to scour it for B-modes. But they barely had to scour. The B-modes were glaring.

They couldn’t figure out what they’d done wrong. They could have sworn they’d accounted for any spurious polarization, any stray morsel of heat. The detectors had passed every last performance test with flying colors. Where was this thing coming from?

They split the data in half, made a map from the first year and a half of observation and a map from the second year and a half. Then they subtracted them. They figured if the signal went away, they’d know it had been in both halves equally, that it hadn’t changed over time. But if it had changed over time—well then it wasn’t cosmological, it was an engineering blip. They ran the test. The signal canceled out. It wasn’t a blip.

They split and recombined the data in every which way, pushed themselves to imagine even the most unlikely scenarios that would have the signal originating in the instrument. Again and again they came up empty handed. Eventually there was no alternative left standing: the signal was coming from the sky.

Of course, there was always the issue of the dust. Everyone knew that interstellar dust in the Milky Way could polarize the photons and mimic the effect of gravity waves. Obviously the dust contributed to the signal, but the question was, how much? The Southern Hole at 150 GHz ought to be pretty clean. That's why they chose it. But you never know.

Obviously dust contributed to the signal, but the question was, how much?

The team didn't have access to any full sky maps with a decent signal-to-noise of polarized emissions from dust—but they knew exactly who did. The ESA’s Planck satellite had been mapping the dust from space and ought to be able to tell them exactly how much of it was contributing to their signal. The BICEP2 team submitted a request to share data. Request denied. They waited, then tried again. Request denied. Was the Planck team being competitive or did they simply feel the data wasn’t ready? Who could say. Either way, Kuo and his team were simply going to have to make do with whatever data they could get their hands on.

POLENIGHTSKY1 — As the Milky Way passes overhead, charged particles of the aurora australis billow over the Dark Sector.

They combed the literature for the leading dust models and fed the results of five of them into their own model. Unfortunately, the models were all built from observations of unpolarized dust at various points on the sky, which were then extrapolated. But without Planck’s actual data, it was their best shot.

They used the models to create fake maps of dust, and they put in 3 million CPU hours on the Harvard supercomputer simulating the results 500 times. The signal wasn't going away. Even after they subtracted the signal for the dust, the B-modes appeared to be still sitting there in plain sight.

That’s when they noticed that a member of the Planck team, J.P. Bernard, had given a public lecture on the dust data. His presentation contained a slide with an image of the dust map. The BICEP team figured it was time to get creative. They digitized the image, reverse engineering it to extract their best guess at the raw data. They knew it was an uncertain procedure, but that was ok—they weren’t staking their claim on it. They were just going to use it as model #6.

Again they subtracted the dust, and again the B-modes remained visible, bright as day.

They had to strike the right balance between being careful and being quick. A signal this bright—someone else was bound to see it. They could feel the competition nipping at their heels. They all agreed to not say a word about it to anyone. Not until they were sure. They were at three-sigma certainty—that meant there was a 1 in 740 chance that the signal was a statistical fluke. In physics, three sigma is considered evidence. Five sigma, a 1 in 3.5 million chance…well that’s a discovery.

BICEP2-b-mode — The B-mode pattern from BICEP2

For a year they sat on the result. Kuo was hoping to hell it was real, though if you asked him to bet on it, he wouldn't risk the money. He had a nagging fear that the B-modes were nothing more than mathematical contamination, just mundane E-mode polarization leaking out. The problem was that BICEP2 had only studied a small patch of sky. Each fragment of data is just a little line segment—it's only when you look at the way those lines are drawn across the entire sky that a pattern emerges. If the line segments form a series of symmetric shapes, like circular ripples, that’s an E-mode: the standard pattern produced by the same old density fluctuations that create the hot and cold spots in the CMB. But if the pattern looks asymmetric, like pinwheels turning in a given direction, that's the jackpot. Only primordial gravity waves can turn those pinwheels.

They had data from a 20° patch of sky, which is to say, not a lot. What do you do with the line segments out toward the edges? You see hints of pattern there, perhaps a slight arc, a suggestion of a pinwheel. But what if it's a circle? Your statistics start to break down. So you throw away some signal, a sacrifice to the gods of error bars. But how to strike just the right balance between signal and certainty was far from clear.

One evening in Stanford, after he'd had dinner and helped put the kids to bed, Kuo noticed an e-mail from his grad student, Tolan.

Two years earlier, Kuo had urged Tolan to find a better way to distinguish the E-modes from the B-modes out at the edges. Tolan began working on the problem on the side, “off pipeline.” They were told again and again, stick to the pipeline, it's the only way to keep things running smoothly, and it was. Everyone treated Tolan's work as a kind of side hobby, so he just kept at it, posting updates now and then to the team's internal website.

Kuo opened the e-mail. I've got a preliminary posting of the matrix estimator. Tolan had done it. He had found a way to cleanly separate the B-modes from the E-modes, and he had run their data. Kuo prepared himself for disappointment. He was sure the signal had disappeared. He clicked on the link to the internal website and scanned Tolan’s results.

The signal hadn't disappeared.

The signal had gotten stronger.

The error bars had shrunk, and the certainty had risen—from three sigma to five sigma. A discovery.

That night, the sun went down, but Kuo couldn't sleep.

In the morning he e-mailed Tolan: If this signal is real, this is the home run of all home runs…

If it were real, it would be the closest anyone had ever come to seeing the beginning of time. It would be the smoking gun proof of inflation. It would be a direct look at the quantum mechanical underpinnings of the universe, probing physics at energies a trillion times greater than what particle physicists could achieve in the hallowed tunnels of the LHC. If it were real, Kuo could finally tell his ten-year-old self the answer: if the universe isn't infinite, it is really damn big.

Funny, the difference between experiment and theory. Theory is the stuff of great drama, littered with “aha” moments. It’s Archimedes shouting, “eureka!” in the bathtub, it’s Guth writing, “spectacular realization” in his notebook, it's Linde waking his wife to tell her, “I think I know how the universe was created.” But experiment—experiment is more like life. It's messy and it happens gradually after a good amount of soldering and shivering and the turning of screws. Sometimes the results are null—and sometimes the results are dust—but little by little it adds up to something tangible and true.

Never Again?

Linde and his wife were packing their things for a Caribbean vacation.

They needed it. They’d been working together again, writing paper after paper, producing a whirlwind of work. Linde couldn’t believe how much they’d done. Every time he had a good idea, he was convinced it would be his last.

As people, and as physicists, they were a perfect match. Where Linde had physical intuition, Kallosh had mathematical intuition. What was difficult for one came easy for the other. They saw the universe differently, and while the process was painful, they each raised the other up in their thinking. Not that it seemed so grand in the moment. Every time they were finishing yet another paper, they’d end up shouting, “Never again!” But they’d take a break, perhaps a vacation, and then they’d start all over again. That’s just how it was in their household. Ideas were nourishment. Physics was air.

Linde thought back to his younger days. It was funny now to think he’d ever wondered exactly what he ought to be. Now he understood that he was a theorist for the same reason an artist is an artist or a poet is a poet—because it’s too painful not to be.

At The Door

Kuo walked up the long driveway, the cameraman keeping pace behind him. For Kuo, the B-mode measurement was a technological achievement, the end of a marathon, the feeling of knowing that he had played an indelible part in the grand unfolding of science. But he knew that for Linde it would be something different: a moral victory, the triumph of reason and intuition, a validation 30 years coming. He was itching to tell him, he was rehearsing it in his mind. Five sigma. Clear as day. R equals 0.2. He raised his hand to knock on Linde’s door.

Epilogue

As of October 2014, maps made by Planck suggest that there is far more polarized dust in the Milky Way than theoretical models had predicted and that the entire B-mode signal measured by BICEP2 may be due to dust . Physicists and astronomers still need more data to determine the source of the signal and to figure out whether gravity waves are lurking behind the dust. Kuo is gearing up to head back down to the South Pole in December to set up BICEP3. The new instrument’s field of view will be three times larger than BICEP2’s and will measure light at a frequency of 95GHz. By comparing its results with BICEP2, Kuo and his team say they will be able to differentiate gravity waves from dust. As for Linde, he is hard at work incorporating inflationary theory into theories of fundamental physics, satisfied that the experimental evidence for inflation is overwhelming even in the absence of gravity waves and motivated, as ever, by the theory’s explanatory power and beauty. Science carries on.

Photo credits: NASA/ESA, Anthony Turner/JPL, Steffen Richter/Harvard University, Keith Vanderlinde/NSF, BICEP2 Collaboration

Articles

From Discovery to Dust