After three years of downtime for crucial upgrades, the world’s gravitational-wave observatories were supposed to restart, more sensitive than ever to the tiniest deformations in spacetime.

But technical problems scuttled that optimistic scenario and assured a tepid start on May 24. Expectations for the 20-month campaign, known as Observing Run 4 (O4), are now tempered. Italy’s Virgo detector is not reopening immediately; instead it will undergo repairs to fix damage caused by a broken glass fiber. Additionally, Japan’s new Kamioka Gravitational-Wave Detector (KAGRA) is far from its intended sensitivity and will observe for just one month before shutting down for troubleshooting, with an eye toward restarting in late 2024. This leaves only the U.S.’s two National Science Foundation–funded Laser Interferometer Gravitational-Wave Observatory (LIGO) detectors in Hanford, Wash., and Livingston, La., to emerge fully functioning from the shutdown.

And yet, despite all these setbacks, the scientific potential for O4 is still enormous. The field of observational gravitational astronomy is so new that adding any discoveries to the fewer than 100 detected mergers between black holes and neutron stars is a genuine advance. “We’re really looking forward to getting back to observing,” says Patrick Brady, an astrophysicist at the University of Wisconsin–Milwaukee and LIGO’s spokesperson. Even without extra help from Virgo and KAGRA, LIGO’s upgrades alone should ensure the project discovers nearly 300 new events in O4—roughly one every other day.

The vast majority of these new detections will be gravitational waves emitted by binary black hole mergers, which are invisible to conventional telescopes. At least a few will come instead from seismic crashes between two neutron stars (or between a neutron star and a black hole) that release not only gobs of gravitational waves but also conspicuous amounts of light—kilonovae. In August 2017 Virgo turned on just weeks before a remarkable series of gravitational waves, designated GW170817, propagated through Earth. Within hours, researchers were able to combine Virgo and LIGO data to single out a sliver of the sky where the signal originated.

They alerted other astronomers, who pointed optical telescopes at the spot and, for the first time, saw a kilonova’s light. Detecting more kilonovae can help researchers answer fascinating fundamental questions about the origin of elements, the limits of neutron star mass and even the rate at which the universe is expanding.

Answers from kilonovae may have to wait until Virgo finishes repairs. With just LIGO’s two detectors running, high-confidence detections of most mergers become much more challenging, and in many cases, localizing gravitational-wave signals to a searchable area could be impossible. (As this story went to press, LIGO announced that it detected a new neutron star–black hole merger in its prerun mode. It was unable to precisely localize the source, however.)

LIGO-Virgo-KAGRA (LVK) researchers aren’t the only ones with a stake in O4’s results. Other scientists are also eagerly awaiting fresh data. “I’m still excited.... There’s so many things that we can clarify if we just have a few more detections,” says Lieke van Son, an astrophysicist at the Center for Astrophysics | Harvard & Smithsonian. She is eagerly awaiting more insights about the demographics of merging black holes—specifically, about the range of masses they display. Visualized as a chart, the gaps and peaks of black hole masses can reveal new details about the ways giant stars die.

Scientists outside the LVK collaboration will likely have to wait. Besides the relatively rare detection of events that demand immediate optical follow-up such as ones that produce kilonovae, the LVK collaboration plans to hold off for about 18 months after the first half of the run concludes—a total of 28 months from now—before it updates its gravitational wave catalog and makes the full O4 data public.

On the other hand, real-time notifications with limited data are available to anyone via smartphone apps—now with customized alert sounds.

How to Train Your Gravitational-Wave Detector

No fluid forms gravitational waves. The waves are themselves a distortion of spacetime, a wake caused by the movement of massive objects such as black holes or neutron stars. When gravitational waves wash through Earth, they are almost imperceptible—but not quite.

Detecting gravitational waves requires a finely tuned instrument that is not so different, in principle, from a finely tuned violin. In practice, researchers shoot laser beams down two perpendicular tubes, each several kilometers long. Each beam bounces between mirrors 300 times until both beams recombine near their common source, where they can be analyzed for any trace of gravitational waves.

The waves sweep through, repeatedly altering the length of a laser beam’s path by tiny distances—about one ten-thousandth the diameter of a proton. The infinitesimal expansion and contraction occurs at a frequency of about 10 to 100 hertz and creates a measurable offset between one beam and its perpendicular twin.

How does this happen? Imagine a violinist playing one string with vibrato. By rapidly rocking a finger back and forth, the violinist expands and contracts it to create periodic changes in pitch. A keen-eared listener can easily hear the rate of vibrato. Similarly, gravitational waves expand and contract space to give a laser beam vibrato. Because the laser’s vibrato is much subtler than a violin’s, to “hear” it, researchers have to contrast the varying beam with its steady twin. This allows them to record the frequency of passing gravitational waves, converting laser vibrato into an audible, iconic “chirp.”

Accurately detecting that subproton-length change of the laser requires the instruments to be isolated from noise—random environmental fluctuations of all kinds, be they seismic rumbles or the pecking of ravens.

LIGO and Virgo are so sensitive, in fact, that they must account for the noise produced by the impacts of photons from the laser. “When the photons reflect, they provide a momentum transfer to the mirror,” explains Albert Lazzarini, a physicist at the California Institute of Technology and deputy director of LIGO. To limit noise, the photons must all have the same phase—essentially, they need to all hit the mirror at about the same time. Previously, LIGO and Virgo had only been able to “squeeze” light into the same phase at high frequencies above 50 Hz. Thanks to upgrades, which included installing a 300-meter tunnel at the LIGO sites, the detectors can now squeeze light down to 30 Hz.

That may not sound like much, but low-frequency squeezing has dramatically improved LIGO’s sensitivity, which is measured as a distance to the farthest detectable binary neutron star merger. LIGO was previously sensitive out to 330 million light-years. It has now improved its range by about 50 percent to plumb cosmic depths of about 500 million light-years. (Because this linear distance applies to a three-dimensional volume of space, the sensitivity gain should actually triple the overall rate of detections.)

At Virgo, upgrades were going well until last November, when researchers installed a noise mitigation system on a mirror. They soon discovered that one of the delicate glass fibers suspending the mirror had snapped. A similar incident happened in 2017, when a glass fiber broke, causing delays. For both mishaps, Virgo researchers identified a likely common cause: during upgrades, particles of dust settled on the fibers and weakened them.

“We modified our vacuum chamber in order to protect the fibers, but there is still a small chance that this can happen,” says Gianluca Gemme, a physicist at the National Institute of Nuclear Physics in Italy, who was recently appointed spokesperson of Virgo. After the November 2022 incident, Virgo researchers replaced the broken glass fiber. Excess noise still remained, however. In late April the noise-plagued detector could listen only out to about 88 million light-years—that is, half the sensitivity of its previous observing run. Gemme says the trouble probably stems from a magnet that is “not well attached,” which was at least partly caused by last November’s misbehaving glass fiber. On May 11 Gemme and the Virgo collaboration announced that they would continue repairs rather than try to restart at suboptimal sensitivity.

“This was a difficult decision for us because we have been working for this upgrade since 2019,” Gemme says. “We really want to improve the sensitivity of the detector, understand the noise and solve the problems.” He says the repairs could be complete by the end of June—and adds that a more pessimistic timeline would have Virgo rejoining LIGO in the fall.

Virgo is not the only detector to struggle. Three years ago experts optimistically predicted that KAGRA would be sensitive out to about 424 million light-years. Its first-of-its-kind cryogenic design and underground location were supposed to offer extra protection against environmental noise. But various setbacks have held KAGRA to less than 1 percent of that hoped-for target—the facility’s current sensitivity only extends out to about three million light-years, meaning it can detect gravitational waves solely from sources within or just outside of our own galaxy. According to Jun’ichi Yokoyama, a physicist at the University of Tokyo and KAGRA’s spokesperson, for O4, the detector will first observe for just one month before shutting down to undergo additional commissioning aimed at boosting its sensitivity by a further factor of 10—a value that would still fall far short of its original goal.

“They’re building something no one’s ever built before, and so they’re discovering new problems,” Lazzarini says. “It’s very hard.” KAGRA has been making progress on some problems. The detector’s cryogenically cooled mirrors were plagued by thin layers of frost that rendered them essentially unusable. During the shutdown, KAGRA researchers developed a five-step cooling strategy that has allowed them to chill the mirrors down to a frigid 20 kelvins without forming frost. Other problems remain unsolved. The mirror’s angular sensing mechanism is unstable, for example, and there are still “mystery noises” from as-yet-unknown sources.

“That is the most difficult part of the business,” Yokoyama says, “identifying the origin of the noise and just making strategies to remove them one by one.” He declined to share the team’s analysis of KAGRA’s mystery noise—as well as what possible fixes are being considered to bring the detector back up to its projected sensitivity.

All the Gravity We Can Hear

By 2021 gravitational waves had so expanded the catalog of known black holes that the LIGO team issued a challenge on Twitter: come up with a collective noun for describing black holes in bulk. Onlookers suggested hundreds of candidate nouns (“a crush,” “a void,” “a scream,” “a disaster,” “a mass,” and so on), though there was no official winner.

Consider this a late submission: A “burden” of black holes—a play on both their physical mass and their solemn existence—carries great weight for physicists. Whatever you might call them, the “burden” black holes place on the LVK collaboration and O4 is a task scarcely changed from all the runs that came before: to chip away at the unknown, adding one newfound merger at a time to gradually glimpse any larger trends that emerge.

Already, it’s clear that most of the black holes seen (or rather heard) via gravitational waves formed from the supernova death throes of massive stars. Such stars burn brightly and briefly, collectively shaping entire galaxies through stellar winds and, near their end, fusing lighter nuclei into heavier elements such as oxygen and aluminum that can serve as feedstock for future generations of stars and planets. Yet these oh-so-important massive stars are also relatively rare, sufficiently sparse in our galactic neighborhood to stifle detailed studies of their stellar evolution. The LVK collaboration’s burgeoning burdens of black holes offer a new avenue of inquiry. “All of the [black hole] fossils of these massive stars give us a way to do archaeology on how massive stars lived and learn what their lives were like in a way that we never could before,” van Son says.

Patterns are starting to form in the mass distribution of black holes. A low-mass peak of around 10 solar masses, under which there are very few black holes, suggests, for instance, that most come from moderately heavy stars.

Theorists predict another peak should appear around 45 to 65 solar masses because of “pair instability” supernovae. These are more powerful than run-of-the-mill supernovae, which occur when a star runs out of fuel and implodes with a sudden loss of radiation pressure akin to demolishing the supports holding up a roof. Pair-instability supernovae arise from very massive stars, 130 to 250 times the mass of the sun, largely bereft of heavy elements—a situation that can accelerate their demise well before they run out of fuel. The thermonuclear fusion within can become so energetic that instead of holding up the outer layers, the photons generated in the star’s core spend their energy creating matter-antimatter pairs, which can accelerate its implosion.

Theorists are now eyeing a spike in the black hole distribution at 35 solar masses as a potential, lighter-than-expected pair-instability peak—or instead a sign of something totally different. “I think we’re looking at a completely unexplored, unexplained feature,” van Son says.

Other features are also predicted but as yet unseen in the black hole mass distribution. Theorists expect O4 could reveal a cliff around 60 or 70 solar masses because such heavyweight black holes would require massive stars that would be too unstable to shine. If no such cliff appears, it could indicate the existence of another undiscovered route to bulky black holes.

More subtle sources of gravitational waves should lurk in the universe as well. And there is a chance—slim, in part because of suboptimal detector sensitivity—that the LVK collaboration could suss them out in O4.

While mergers between black holes are short-lived events, some sources should produce continuous gravitational waves. As detectors grow more sensitive, they could pick up on the gravitational-wave background—a kind of white noise formed by eons of the waves washing back and forth. Or they could hear a “mountain” on a neutron star. When the mass of several suns is compressed into an orb that is 10 km in diameter, extreme gravity crushes mountains of superdense material into molehills that are millimeters tall. These mountains contain plenty of mass—a teaspoon of neutron star matter weighs 10 million metric tons. A rapidly rotating neutron star could whip these mountains around so fast that they’d create regular ripples in spacetime. As heard by a gravitational-wave detector, the waves would not crescendo to a chirp but stay a continuous hum. Whether this is detectable during O4 is “a question of how big the mountain is on the surface of the neutron star,” Brady says.

Luck looms large over discoveries. The spectacular binary-neutron-star-produced GW170817 from August 2017 was visible to astronomers because a jet of energetic light happened to point at Earth. But basic geometric reasoning shows that in 90 to 95 percent of cases, such jets should be pointed too askew from our planet to be detectable. Other estimates suggest only about six to 10 neutron star events should occur in O4. That is a decent chance but no guarantee that one will be aligned to offer another direct view of its jets—especially when technical troubles have sidelined two out of three gravitational-wave projects, making each event’s localization on the sky a far higher hurdle.

The LVK collaboration’s goal for neutron star mergers in O4 is to catch them as early and often as possible, giving other observatories more opportunities to rapidly perform follow-up investigations. When and if O4 detects a promising signal that can be localized (unlike the just announced prerun detection), within 30 seconds, the LVK collaboration will send out an alert to a consortium of optical telescopes. “We want to be able to send these alerts out as soon as possible so the astronomers can actually start pointing their telescopes and be ready,” says Surabhi Sachdev, a physicist at the Georgia Institute of Technology and a LIGO member.

Although the LVK collaboration can’t tell other telescopes what to do, its members hope that, given an enticing enough event, astronomers will choose to try and spot the source of whatever gravitational-wave detectors have heard.

Listening for the Future

Even if Virgo and KAGRA were fully functioning, truly novel science would have to wait. Next-generation ground-based detectors, such as the planned Einstein Telescope and the Cosmic Explorer, are intended to be so sensitive that they would hear every binary black hole collision in the observable universe. Roughly every second, they would detect a new merger.

Such numbers would wholly invalidate proprietary data policies, such as the LVK collaboration’s current practice of keeping most of its finds private for about a year and a half. “The setup that they have right now is not sustainable,” van Son says. As such, the plan is that the next generation of detectors will release data immediately, for anyone to analyze.

“They will really be a giant leap forward,” Gemme says. “But this leap will be made possible by all the things that we are learning in the current detectors.” KAGRA’s troubles, for example, could translate into know-how that will help researchers navigate the even more ambitious cryogenic systems planned for the Einstein Telescope.

If they work as planned, Cosmic Explorer and the Einstein Telescope will usher in a new era of astronomy: one where gravitational waves are no rarer than radio waves, and faint signals—from collisions in the early universe to molehill-sized mountains on nearby neutron stars—echo around the world, loud and clear for all to hear.

Editor’s Note (5/23): This article was edited after posting to correct the descriptions of the kinds of stars that are likely to lead to black holes and the thermonuclear fusion that occurs in pair-instability supernovae. The text had previously been amended on May 23 to correct descriptions of kilonovae.