Suppose you want to glimpse the beginning of time, the very first moments of cosmic creation. You might start by building a perfect telescope, an instrument so powerful that it could see to the far end of the observable universe. You'd scout out a dry mountaintop, far from the star-fading glow of civilization. You'd level out a perch near its peak and place a state-of-the-art observatory atop it. You'd outfit it with a gigantic mirror—something much larger than could be launched into space—and equip it with a series of sophisticated detectors. You'd spend several years and several billion dollars, so that every last photon was within your reach. But what could you see with it? Say it was that one night in an astronomer's thousand, when the moon hides below the horizon, and the sky appears as a clear, dark dome overhead. What jewels would glitter out from that purplish-black showcase of celestial sights?

Quite a few, it turns out. In the foreground, you would see a smattering of planets, their orbits adrift against the fixed whirl of the constellations. Beyond them, local stars would loom large against a backdrop of fainter specks of white. In the sky's darker corners, galaxies would glow, some from hundreds of millions of light-years away. If you pointed your perfect telescope at exactly the right spot, it could reveal deeper cosmic recesses still. It could take you to the very first stars—the huge hydrogen and helium spheres, whose fiery surfaces illuminated the young universe.

But light has limits; it can't show you the entire universe. You could look through a telescope all night, every night, and never see into the center of a black hole or back to the dawn of time itself. For the first few hundred thousand years after the big bang, photons of the infant universe stayed trapped in a dense soup of light-suffocating particles, like fireflies sealed into sludge. It was not until 380,000 years after the big bang that the universe cooled into something transparent and, for our purposes, legible—a void through which the flash of creation could be seen. We call this flash the cosmic microwave background (CMB), and it is the dominant text of modern cosmology. It is also a wall, a barrier in time, beyond which darkness reigns.

For centuries now the careful collection of ancient light has been the dominant way to observe the universe, the key to cosmology's most ambitious experiments. But light cannot illuminate the beginning of time, no matter how large and sophisticated our telescopes grow. To see beyond the CMB, back toward the dawn of the universe, cosmologists must turn to gravity, a force that leaves echoes of its own strewn across space—echoes we call gravitational waves. To detect these echoes, we will need a new kind of instrument, something very different from a telescope.

The First Detectors
The quest to build an instrument that can detect gravitational waves began decades ago, but so far it has proved fruitless. As of this writing, LIGO, the $570-million Laser Interferometry Gravitational Wave Observatory, represents the best such attempt [see “Ripples in Spacetime,” by W. Wayt Gibbs; Scientific American, April 2002]. It consists of three instruments, two in Washington State and one in Louisiana. Each of these is an engineering marvel, a laser-based measuring stick capable of detecting a twitch the width of an atom. LIGO works by shooting laser beams down two perpendicular arms and measuring the difference in length between them—a strategy known as laser interferometry. If a sufficiently large gravitational wave comes by, it will change the relative length of the arms, pushing and pulling them back and forth. In essence, LIGO is a celestial earpiece, a giant microphone that listens for the faint symphony of the hidden cosmos.

Like many exotic physical phenomena, gravitational waves originated as theoretical concepts, the products of equations, not sensory experience. Albert Einstein was the first to realize that his general theory of relativity predicted the existence of gravitational waves. He understood that some objects are so massive and so fast moving that they wrench the fabric of spacetime itself, sending tiny swells across it.

How tiny? So tiny that Einstein thought they would never be observed. But in 1974 two astronomers, Russell Hulse and Joseph Taylor, inferred their existence with an ingenious experiment, a close study of an astronomical object called a binary pulsar [see “Gravitational Waves from an Orbiting Pulsar,” by J. M. Weisberg et al.; Scientific American, October 1981]. Pulsars are the spinning, flashing cores of long-exploded stars. They spin and flash with astonishing regularity, a quality that endears them to astronomers, who use them as cosmic clocks. In a binary pulsar system, a pulsar and another object (in this case, an ultradense neutron star) orbit each other. Hulse and Taylor realized that if Einstein had relativity right, the spiraling pair would produce gravitational waves that would drain orbital energy from the system, tightening the orbit and speeding it up. The two astronomers plotted out the pulsar's probable path and then watched it for years to see if the tightening orbit showed up in the data. The tightening not only showed up, it matched Hulse and Taylor's predictions perfectly, falling so cleanly on the graph and vindicating Einstein so utterly that in 1993 the two were awarded the Nobel Prize in Physics.

The trouble for LIGO is that it can hear these binary pulsars only in their final moments, when their starry spiral accelerates, churning out a series of strong waves that propagate across space like an invisible cosmic death rattle. Our universe may be large and star-filled, but binary collapses are rare. To hear them with any regularity, you have to train your ear on a gigantic chunk of the cosmos. Until recently, LIGO's reach was limited to a region of space that can go centuries without a binary collapsing within its borders.

But LIGO's first build was a dry run, a way of working out the engineering kinks that accompany instrument integration on a kilometers-wide scale. Now that LIGO's engineers know they can make a complex detector work, they are upgrading its sensitivity, so that soon it will be able to detect a binary collapse from 500 million light-years away—an improvement that could allow it to hear hundreds of these events a year. Indeed, most astrophysicists expect LIGO to achieve the first direct detection of gravitational waves within months of its return in 2016—the 100th anniversary of Einstein's prediction.

Atom Waves
Despite its considerable cost, LIGO's ambitions are limited. In some ways, it is a proof-of-concept mission, a necessary first step before gravitational-wave science ascends to its most natural environment: space. Our planet is a terrible place for a gravitational-wave observatory because its crust is constantly awash in seismic noise—the product of booming tectonic collisions underneath Earth's surface and sloshing oceans atop it. All of this shaking and quaking can easily drown out the thin, matter-shifting wisp of a gravitational wave. To hear a wider variety of them, we need a detector in the abyss beyond the atmosphere, where conditions are considerably more serene.

At the NASA Goddard Space Flight Center, two teams of engineers are positioning themselves to be the first to put a gravitational-wave detector in space. The older of these teams has been refining its mission, the Laser Interferometer Space Antenna (LISA), for decades. The LISA mission is an audacious engineering project, demanding a level of precision that makes LIGO look Lego-like by comparison. It requires the launch of three spacecraft that orbit the sun in the form of an equilateral triangle with sides five million kilometers long. Once the spacecraft are in place, the distance between them will be measured, continually, with lasers. If a gravitational wave rolls through, disturbing the spacecraft and distorting the triangle, the lasers will capture it.

LISA's basic design has not changed much since a few pioneers of gravitational-wave science sketched it onto a cocktail napkin at a NASA physics conference more than three decades ago. But it has grown refined over time, as engineers have grappled with the practical challenge of bringing its ambitious design to life. In the late 1990s and early 2000s, LISA emerged as an early contender to become NASA's next flagship astrophysics mission, following the James Webb Space Telescope (JWST). But in the years since, the JWST has swallowed most of NASA's astrophysics budget, and with no detections at LIGO, astronomers have found it hard to make a case for a multibillion-dollar gravitational-wave detector. A green light for a mission such as LISA could be more than a decade away.

These delays have created space on NASA's drawing board for novel ideas about how to detect gravitational waves in space. A small team within the agency's Advanced Concepts division recently began developing a new kind of gravitational sensor, based on a nascent technology called atom interferometry. The team is loosely organized, and so far its work can hardly be said to constitute a full-blown mission. Its principal leaders—Babak Saif, an interferometer engineer for the JWST, and Mark Kasevich, a professor of applied physics at Stanford University, are both engrossed in other pursuits. This is a side project for them, something to tinker with and dream about in the margins of their workweek.

In February, I visited Saif at one of Goddard's laser labs, where he is slowly starting to build an atom interferometer, a technology that he expects to form the basis of a smaller, more nimble gravitational-wave detector. As one of the world's most prestigious space research labs, Goddard is home to a slew of scientists with gaudy academic pedigrees, but Saif had humbler beginnings. After immigrating to the U.S. from Iran at the age of 17, Saif's family settled in northern Virginia, where he began taking classes in science and mathematics at a local community college. Saif worked nights at a gas station to support himself and proved himself a quick study at school. In 1981 he transferred to the Catholic University of America on a full scholarship, and in the years since he has completed two Ph.D.s. Before coming to Goddard, Saif spent a decade at the Space Telescope Science Institute, where he designed the interferometer that will eventually test the mirrors of the JWST. Saif's interferometer will ensure that the mirrors are accurate to the nanometer scale to avoid a repeat of the fiasco that befell the Hubble Space Telescope when it reached orbit with a misaligned mirror.

Saif explained that his and Kasevich's mission concept is similar to LISA's in that it involves measuring the distance between orbiting spacecraft. But whereas LISA measures changes in distance by combining light from the laser beams shot between the spacecraft, Saif and Kasevich's mission will instead employ atoms sitting just outside the spacecraft. Because the atom interferometer measures distances between atom clouds, not spacecraft, it can be much smaller. Its current design calls for arm lengths that are 5,000 times shorter than the LISA design.

The power of this technique is in its precision. A gravitational wave might shift the distance between the spacecraft by less than a trillionth of a millimeter, and yet the atom interferometer will detect the difference.

Not everyone is enthused about atom interferometry, however. The limited funding that exists for space science has led to tension between Saif's atom interferometry team and the LISA team. The two mission concepts are similar in some ways. Both require precision coordination between spacecraft, and both make use of interferometry to make precise measurements. But according to Saif, the switch from light interferometry to atom interferometry will allow for a cheaper and more sensitive detector and a reduction in the enormous distance between spacecraft; the latter has long been a sticking point for LISA's critics.

The LISA folks fire back by attributing the cost savings of atom interferometry to its newness. They point out that wide-eyed boosters of new technologies often underestimate the heavy costs of development. A design's real price tag emerges only once a mission is in place, they say, because only then do you begin to see the more difficult engineering challenges that come with system integration.

The Trouble with Light
At Goddard, I asked Saif what motivated him to spend his spare time on such a speculative mission, one that may never fly. He told me it was the possibility of new physics that fascinated him. He said he expects the next few decades to usher in an epochal transition in the field of astronomy—a switch from the photon to the graviton.

Indeed, gravitational waves help to make up for a host of light's scientific liabilities—and not just its inability to tell us about the beginning of time. Light has other limitations as an information carrier. To start with, it is the product of interactions between particles. When light springs out into the universe, it announces the occurrence of tiny events, such as the fusing of hydrogen into helium inside of stars. It is a record of the infinitesimal. If we want to learn how large objects move through spacetime, we have to aggregate light from scores of these tiny events and use it to make inferences. We have to piece together a surface-layer mosaic.

Worse still, light biases our view of the cosmos because it tends to come from thermodynamically intense environments. Astronomy's large, signal-worthy splashes of light are the products of fiery events, such as stars in their supernova death throes. When we summon the universe to mind, the structure we see is slanted toward hot, chaotic places.

Light signals are fragile, too. They often dilute or disappear altogether, as they make their way across the cosmos. Some are absorbed by giant gas clouds in their path. Others scatter or fall into deep gravity wells, never to be heard from again. The deepest of these wells are supermassive black holes, the pillars of cosmic structure around which entire galaxies pivot. Scientists want to know more about these black holes—especially what happens when two of them merge together—but no light from a black hole ever reaches our telescopes or eyes because photons, speedy as they are, cannot escape the suction of a black hole's center.

Instead cosmologists have to content themselves with light that a black hole does not devour, light that springs out from its periphery, from matter caught in the furious distortions of spacetime around it. Luckily, gravitational-wave signals aren't nearly as impressionable as light. They don't scatter or dilute. Instead they ripple through the universe cleanly, impervious to the astrophysical giants in their path.*

Primordial Echoes
A few weeks after my trip to Goddard, I visited David Spergel, chair of Princeton University's astrophysics department and one of the world's preeminent cosmologists. Spergel chairs the National Research Council's decadal survey committee on cosmology and fundamental physics, the reports of which play a large role in determining the long-term research priorities for cosmology. NASA is known to pay especially close attention to its recommendations, which means Spergel has an outsized say in what science missions the agency decides to fly.

As we sat down in his office, Spergel began detailing the advantages of gravitational waves. Unlike light, he explained, the universe has always been transparent to gravitational waves. There was no primordial era during which they were hidden by strange cosmic conditions. Indeed, gravitational waves would have no trouble rippling out to us from the very first moments after the big bang. But how do we know any were around then?

“To produce gravitational waves, you have to move a lot of matter around very quickly, and one way you could do that is with a phase transition,” Spergel told me. A phase transition occurs when a physical system changes states. The classic example is water freezing into ice, but there are also cosmic-size phase transitions, some of which occurred shortly after the big bang. Take quarks, for example. Today quarks are mostly bound up in the nuclei of atoms, but in the first microseconds of the universe, they buzzed around freely in what cosmologists call a quark-gluon plasma. At some point, the universe transitioned from this quark-gluon plasma to a new phase populated by protons and neutrons.

“If you had a first-order phase transition like that, bubbles would form within the plasma, causing a whole lot of matter to move around quite violently,” Spergel said. First-order phase transitions occur suddenly, when bubbles of a new phase form in the midst of the old one. These bubbles expand and collide until the old phase disappears completely, completing the transition. The chaos of this process would have generated strong sets of gravitational waves, which may be washing over us today. Their detection could offer our first glimpse into the universe's infancy.

And there might be older gravitational waves, still. In some inflationary models of the universe, the first burst of exponential cosmic expansion coincides with quantum fluctuations of spacetime—ripples that cause certain regions of the universe to expand faster than others. These fluctuations could have given rise to a special species of gravitational waves, called stochastic gravitational waves, that would have formed when the universe was less than a trillionth of a trillionth of a trillionth of a second old [see “Echoes from the Big Bang,” by Robert R. Caldwell and Marc Kamionkowski; Scientific American, January 2001].

“Most inflationary models of the universe predict this stochastic gravitational-wave background coming from the very early parts of the universe,” Spergel told me. “If we could observe it, it could show us fundamental physics. It could show us what the universe looked like at energy scales that are 1013 times what we're getting at the Large Hadron Collider,” he said.

Going after stochastic gravitational waves is high-stakes science. Detecting them would be very difficult. It would require an especially sensitive instrument, and painstaking data analysis, to sift out the precious primordial waves from the legions of gravitational-wave signals that would bombard a space-based detector. If you could collect this signal from every corner of the heavens and scrub it of stray noise, you would have a stochastic gravitational-wave background, an all-sky map of gravitational waves. You'd have a new foundational text of cosmology to pore over.

The mission designs for both LISA's and Saif's atom interferometry concepts are aimed at detecting gravitational waves from more conservative targets, such as black hole mergers. In headier days, LISA's designers dreamed up a Big Bang Observatory, a successor mission tuned specifically to stochastic gravitational waves. But such an observatory was always a long shot, an idea that was decades away from implementation. Saif told me he would like to reverse the mission order and go after stochastic gravitational waves first, but so far the designs he has worked up target the same signals as LISA. The conservative approach is a diplomatic sop to the wider community of astrophysicists, who are intrigued by gravitational-wave science but want it to start slow, by targeting objects already known to exist.

“Supermassive black hole collisions are the bread-and-butter work of gravity-wave experiments,” Spergel told me. “If we fly one of these spacecraft and we don't hear huge black holes colliding, then something is very wrong with our picture of the universe,” he said. “But the home-run signal is the cosmology.”

At some point, Spergel's decadal survey committee may find itself choosing between black holes and cosmology and, perhaps, atom interferometry and light interferometry. The committee is set to reconvene midway through the decade to evaluate and adjust the course it set in 2010. By the time the next such survey comes around, the JWST will have launched, presumably freeing up money for an ambitious space science mission.

As I stood to leave, I asked Spergel if he had an early favorite, if he thought Saif's mission would best LISA in the long haul. He told me that he isn't convinced the atom interferometry concept will win out, but he is convinced it's interesting enough to think hard about. Then he told me a story. “Many years ago, well before he won his Nobel Prize, I was talking to Steven Chu about how to do great science, and he told me something I've never forgotten,” Spergel said, walking me out. “He said you have to put yourself in the position to do experiments that could be important,” he continued. “I think both these experiments fall into that category.”

*Correction (10/2/13): The last two sentences in this paragraph are erroneous. Gravitational waves can be scattered or diluted, but only barely, and ripple through the universe cleanly, nearly impervious to the astrophysical giants in their paths.