Euclid telescope to seek source of dark energy—the biggest mystery in the universe | Science

A version of this story appeared in Science, Vol 380, Issue 6651.Download PDF

When the Euclid space telescope blasts off from Cape Canaveral in Florida early next month, it will embark on an unprecedented effort to survey 1 billion galaxies—and perhaps solve cosmology’s greatest mystery. The search will cover more than one-third of the sky and look back in time to galaxies shining when the universe was just one-quarter of its current age of 13.8 billion years. Although the task is immense, Euclid’s primary goal is surprisingly simple. The data it collects will be boiled down to a single number, denoted by w. And cosmologists are hoping, maybe even a bit desperately, that it is not –1.

w describes the effect of dark energy, the mysterious antigravitational force that is accelerating the expansion of the universe. All measures so far suggest that w is close to –1. If it proves to be exactly that, it will confirm the vanilla solution to dark energy: that it’s a simple tweak—a cosmological constant—added to Albert Einstein’s theory of gravity, which bestows empty space with an innate springiness of its own. As the universe expands, giving birth to more space, the total amount of dark energy also grows—so that the energy density always remains constant.

That solution is anathema to cosmologists because it is simply a fudge factor that does not explain where dark energy comes from and why it has that value. “If w does equal –1, we still don’t know what it is,” says astrophysicist Ofer Lahav of University College London. A slight deviation from –1 or, even better, a value of w that changes with time could point cosmologists toward a new, overarching theory that would entail a fresh understanding of physics. “The hope is to find a wrinkle,” says Mark Halpern, a cosmologist at the University of British Columbia.

The €1.4 billion Euclid mission, developed by the European Space Agency (ESA), is not tackling this problem alone. Many smaller studies preceded it and Euclid is the first of several multibillion-dollar dark energy projects that will soon debut, including the Vera C. Rubin Observatory, a U.S.-funded survey telescope in Chile that will open its eye in 2025, and NASA’s Nancy Grace Roman Space Telescope, to launch in 2026. “The whole community is spending billions of dollars to see if w is –1. If it is not, there will be another Nobel Prize,” Lahav says.

At the very least, researchers hope the telescopes will narrow a growing list of alternatives to a cosmological constant, including revisions of Einstein’s theory of gravity and theories invoking a new physical force that could come in various flavors and vary in time. “Every week there are new theories,” says cosmologist Celia Escamilla-Rivera of the National Autonomous University of Mexico. “But this soup of models needs data.” For Jason Rhodes, an astrophysicist at NASA’s Jet Propulsion Laboratory (JPL), it’s an exciting time to be a cosmologist. “We might be on the precipice of discovering new physics.”

It was 25 years ago that a team of astronomers shocked the world with a report that the expansion of the universe set in motion by the big bang was not slowing because of gravity, as everyone expected. Instead, it was somehow speeding up. The researchers relied on measurements of 50 supernovae of a particular kind, known as type Ia: white dwarf stars that explode with an inherently predictable brightness. The team found that the most distant ones were dimmer—and hence farther away—than they should have been if the cosmic expansion was steady or slowing down. “The community was completely confused,” Lahav remembers.



What came to be called dark energy seemed to be at work. Since then, the evidence for it has continued to mount, from hundreds more supernovae, observations of how galaxies clustered over time, and patterns in the big bang’s afterglow, the cosmic microwave background (CMB). A consensus model, called lambda-CDM, has emerged in which the universe has three components: 5% normal matter, mostly atoms; 27% cold dark matter (CDM) made up of some as yet undetected particle; and 68% dark energy from a mysterious source.

The Greek letter lambda in the cosmic recipe represents one possible explanation, the cosmological constant, which Einstein himself had proposed in 1917. He was attempting to fix a problem with his theory of gravity, general relativity. At the time, astronomers believed the universe was static so, without something to oppose it, gravity would pull the universe toward collapse. Adding a constant amount of energy to empty space did the job, but Einstein was never comfortable with the idea. He dropped it a little over a decade later when Edwin Hubble and others showed that distant galaxies were in fact being flung apart in the wake of the big bang.

Now, lambda is enjoying a second life, but it remains a fudge factor without any physical explanation, and one possible basis for it doesn’t hold up. According to quantum mechanics, the theory of the atomic world, the vacuum should be abuzz with energy from fluctuations that cause particle-antiparticle pairs to constantly pop in and out of existence. But the predicted energy in that quantum fizz is much too big: 10120 (that’s 1 followed by 120 zeros) times the value of lambda astronomers have observed. “It’s incompatible with us existing,” says Ue-Li Pen, a theoretical astrophysicist at the Perimeter Institute. “The universe would blow up.”

Graphic: Clashing forces mold the cosmos

The expansion of the universe has been shaped by dark matter and dark energy, both poorly understood. Early on, dark matter’s gravity slowed cosmic growth. Later, as galaxies spread apart, dark energy began to accelerate the expansion. The Euclid mission will chart the changing expansion rate further back in time than ground-based telescopes.

Graph showing the size of the universe over time since the big bang, 13.8 billion years ago. It is shaped like an elongated bell with the open end facing to the right. The size increases exponentially after the big bang, then the growth slows (labeled “deceleration from dark matter”), then slowly begins to increase again (labeled “Acceleration from dark energy”) and is expected to continue increasing past the present day. Two arrows pointing back along the time axis represent telescope reach. Ground-based telescopes reach back about 3 billion years, but Euclid is expected to reach back about 10 billion years.

Probing dark energy

Dark energy has been quantified by studying radiation left from the big bang and mapping galaxies in more recent times. But astronomers want to know what happened in between They rely on three main techniques.

Cosmic distance ladder. Supernovae with predictable brightness provide distances to galaxies. Their Doppler-shifted light shows how fast those galaxies recede, revealing the cosmic expansion rate.

Galactic vibrations. The seeds of galaxies cast ripples in the gas of the early universe. These ripples have grown with cosmic expansion, providing a yardstick to measure it.

A distorting lens. Measuring how dark matter distorts images of distant galaxies reveals the clumpiness of matter and how it has evolved. Dark energy acts as a brake on growing clumpiness.

What lies ahead

Determining dark energy to a level of 1% or better could help theorists explain its origin. Contending theories fall into three main categories.

Cosmological constant. The simplest explanation for dark energy fits current data. The vacuum has an innate antigravitational energy that grows as more space is spawned, driving cosmic acceleration.

Quintessence. If the acceleration is not constant, dark energy could be a fifth fundamental force that is sometimes attractive and sometimes repulsive leading to a variety of fates for the universe.

Modified gravity. If gravity at cosmic scales works differently from in the local universe, the need for dark energy and dark matter can disappear.

(Graphic) C. Bickel/Science; (Data) European Space Agency

Cosmologists describe space as a perfect fluid and w is defined as the ratio of the fluid’s pressure to its energy density. A negative value indicates an outward pressure, and a value of exactly –1 signifies that the pressure is a constant, unchanging feature of the cosmos. A deviation from –1 would point to a dark energy density that is growing or declining with time—and a universe that could end up accelerating even faster, or eventually start contracting. But all estimates so far—from the CMB and the universe’s earliest days, and from its most recent few billion years—suggest that w is close enough to –1 for it to sit within the error bars.

To get a more precise reading, astronomers want to fill in the middle part of cosmic history, in particular the period more than 7 billion years ago when the universe was smaller and gravity still dominated over dark energy. They want to observe the transition from deceleration to acceleration, but have not had the tools to look back that far. Earth’s atmosphere limits ground-based telescopes from making sufficiently precise measurements beyond about 3 billion years into the past. “We have to get this right, and you can only do it from space,” says Euclid Project Scientist René Laureijs at ESA.

In 2007, ESA received proposals for two missions that, using different techniques, would take the search for dark energy into space for the first time. One proposal, called the Dark Universe Explorer, would rely on a technique called weak gravitational lensing: tiny distortions in the shapes of distant galaxies that result as the gravity of intervening matter bends their light. Unlike strong gravitational lensing, where one sees smeared arcs or copies of celestial objects, these distortions are imperceptible to the eye; it is only by statistically comparing large samples of galaxy images that weak lensing can be measured.

Those measurements give astronomers a handle on the clumpiness of matter between the imaged galaxies and Earth. And by making those measurements for galaxies at different distances, they can see how the fight between gravity and dark energy evolved over time.

The second proposal, called the Spectroscopic All-Sky Cosmic Explorer (SPACE), would exploit so-called baryon acoustic oscillations (BAOs). These originated soon after the big bang when the universe was a roiling soup of gas and photons. Denser clumps of matter emitted pressure waves through the soup, concentrating matter in ripples around each clump. These ripples can be detected statistically in the CMB, and today their imprint can be seen as a peak in the distribution of galaxy separations at 490 million light-years.

By mapping millions of galaxies and calculating the size of this standard yardstick at different times, astronomers can chart the expansion rate and see the effect of dark energy. The galaxy distribution “is a cosmological laboratory,” says Andrea Cimatti of the University of Bologna, who led the SPACE proposal. “But to achieve the accuracy needed requires a very large set of data, a very large volume of the universe.”

With two compelling cosmology missions to choose from, ESA suggested merging them onto a single spacecraft and, in2011, Euclid was approved. But it wasn’t an easy marriage. “It was complicated,” says John Peacock of the University of Edinburgh, one of the founding fathers of the Euclid mission. “There were a pile of experimental trade-offs.”

For example, there were competing demands for the light from Euclid’s 1.2-meter main mirror. The weak lensing survey requires visible light to make sharp images, whereas the BAO survey relies on infrared light to map more distant galaxies, whose light is “redshifted” by the expansion of the universe. Euclid managers commissioned optics-maker Zeiss to cast a complex, layered piece of glass that allows infrared light to pass through to one set of instruments while reflecting visible light along a different path to another set of instruments.

Structural and thermal model of the Euclid satellite in a lab.
In a 6-year survey, Euclid’s 1.2-meter main mirror will scan more than one-third of the sky and look back 10 billion years. It aims to record 10 billion objects.Stephane Corvaja/ESA

Engineers also elected to build Euclid’s mirrors and other components from silicon carbide, a material used in car brakes and bulletproof vests, because temperature swings don’t alter its shape—critical if perfect weak lensing pictures are to be taken reliably over a 6-year survey. But silicon carbide is difficult to work with: Components must be molded from a powder and baked into ceramic. A baseplate that holds the telescope’s detectors was especially tricky, says Euclid Project Manager Giuseppe Racca at ESA. Technicians kept finding tiny cracks, and casting a flawless one took 2 years longer than planned. “We had two or three failures,” Racca says.

The BAO survey team had to make some compromises. To measure precise distances to galaxies, researchers need accurate redshifts that can only be gleaned from spectra. The team had wanted to use an array of thousands of tiny switchable mirrors to deflect light from individual galaxies into a light-splitting spectrograph, gathering spectra from many galaxies at once. But ESA deemed the device too risky as it had never been tested in orbit. Instead, the survey relies on a grating prism, or grism, to spread out the light from every source in the field of view and record all the spectra simultaneously with a camera. “It’s simple and brutal,” Peacock says. But that approach requires complex image processing to untangle overlapping spectra and remove spectra from unwanted objects, such as foreground stars.

The detectors for the infrared camera also proved to be a headache. Euclid’s designers wanted to use detectors made by Teledyne Technologies, in California, but because they have military applications, they are subject to U.S. export controls. NASA played the intermediary, acquiring the sensors and carrying out testing and packaging at JPL in return for more than 100 seats for U.S. researchers in the Euclid Consortium, the 2000-strong body that built the instruments and will process the data. But the extra bureaucracy was burdensome. “Even talking to my European colleagues, I needed to get the conversation cleared,” Rhodes says.

By 2022, the spacecraft was almost complete. Then Russia invaded Ukraine and Western sanctions scuttled Euclid’s scheduled launch on a Russian Soyuz rocket. Spacecraft are designed to withstand the launch stresses of their chosen rocket, so switching is not simple. But by late last year, ESA had found a suitable alternative in SpaceX’s Falcon 9. “It was a period of anxiety and uncertainty,” says Bhuvnesh Jain, a cosmologist at the University of Pennsylvania.

Euclid now sits in a Florida cleanroom, ready for its early-July launch and monthlong journey to L2, a gravitational balance point 1.5 million kilometers from Earth where NASA’s JWST space telescope also resides. Euclid’s U.S. rivals, the Roman telescope and the Rubin observatory, will join the dark energy hunt later. The three projects will look at different parts of the sky, at different wavelengths, and with overlapping combinations of the three primary techniques (weak lensing, BAO, supernovae). Roman, with a larger mirror than Euclid, will peer further into the past but over a smaller sky area. The ground-based Rubin will not see as far as the other two, but its optical and infrared survey will be the widest.

Each team wants to be first to solve the dark energy mystery, but they are likely to need supporting evidence from one another to convince the community that any discovery is real, Rhodes says. “Ten years down the road, joint analyses of data from all three—analyzing at pixel level in clever ways—will put the most compelling constraints on dark energy.”

Theorists have reason to hope that the data will quickly challenge the current picture, noting that current observations offer two clues that all is not right with lambda-CDM. The first is a discrepancy in measurements of the Hubble Constant, H0, which describes how fast the universe is expanding. Studies of the CMB allow researchers to calculate that today a galaxy 1 megaparsec (3.26 million light-years) from Earth should be receding at 68 kilometers per second (km/s), assuming the particular mix of matter, dark matter, and dark energy prescribed by lambda-CDM. But when astronomers measure the actual recession velocity of nearby galaxies, using type Ia supernova distances and other techniques, they clock a speed of 73 km/s. Researchers initially assumed it was simple experimental error—these are hard things to measure—but as methods were refined and new techniques added, the error bars shrank but the so-called Hubble tension remained.

A second tension is over an independent parameter called S8, which describes the clumpiness of matter in the cosmic web of galaxies. Again, given the lambda-CDM recipe, measures of clumpiness in the CMB predict a value of S8 that we should see today. But the nearby universe actually appears to be 10% less clumpy than the prediction. “You should be able to draw a line between the two but when you draw the line, it doesn’t match up,” Rhodes says. The gap is another chink in the armor of lambda-CDM.

To resolve the Hubble tension, theorists have concocted a raft of models, collectively known as “early dark energy.” They suggest that a form of dark energy gave the universe an early push before the creation of the CMB. This would raise the value of H0 derived from CMB data. Escamilla-Rivera considers it a “very good candidate.” Recent observations of the way the ancient light of the CMB is polarized—a tendency to vibrate in certain directions—have hinted at that early growth spurt. But it wouldn’t resolve the S8 tension, and the evidence is nowhere near firm enough to convince Jain. “Dark energy appeared, disappeared, reappeared again—it doesn’t sound pretty,” he says.

Early dark energy, like many other alternatives to the cosmological constant, requires an entirely new force. Some theorists describe the force as a field or fluid pervading the whole universe and give it the name quintessence. The key to quintessence is that it changes in value through cosmic history and can be attractive or repulsive. It might have switched from attraction to repulsion about 10 billion years ago, gradually pushing cosmic expansion toward acceleration. It could be fickle again: It might cause the expansion to grow at a faster than exponential rate, tearing apart the cosmos in a “big rip,” or flip into an attractive force pulling the universe into a big crunch.

Data from Euclid and other dark energy probes could boost this picture. If they can detect even a tiny deviation in w away from–1—evidence that dark energy is not constant in time—that would be a nail in the coffin of the cosmological constant. “The key goal we all want to know is: Does dark energy evolve?” Peacock says.

Another possibility is that the theory of gravity needs a revamp. General relativity solves puzzles near and far, from the orbit of Mercury to gravitational lensing by galaxy clusters, but it has never been tested across truly cosmic distances, where some think it might break down. In addition to measuring the BAO yardstick, Euclid’s spectroscopic survey can chronicle how gravity pulled clusters of galaxies into sheets and filaments over time, forming the cosmic web. If the growth of these structures deviates from the predictions of general relativity, it could indicate that Einstein’s theory itself needs revision, eliminating the need for dark energy. “If Einstein’s theory is incorrect, there’s a huge industry of alternative theories,” Peacock says.

After arrival at L2 and a few months of calibration, Euclid will begin its survey, covering 36% of the sky while avoiding areas crowded with the stars and gas of the Milky Way and looking back 10 billion years. In just 2 days it will survey as much of the sky as the Hubble Space Telescope has in more than 3 decades. It will record an estimated 10 billion galaxies, stars, and Solar System objects, identifying new targets for JWST and large ground-based telescopes to inspect. “It’s a fantastic gold mine for any kind of science,” says astrophysicist Yannick Mellier of the Astrophysics Institute of Paris and leader of the Euclid Consortium.

Out of that flood, the Euclid Consortium will analyze images of 1 billion galaxies for weak lensing and spectra from tens of millions of galaxies to fix their positions for the BAO survey. The aim is to calculate the value of w with an accuracy of better than 1%, an order of magnitude better than current estimates.

Humbled that they haven’t figured out what makes up more than two-thirds of the universe, cosmologists are now collectively holding their breath. “We still don’t know very much. There’s no evidence that it’s not a cosmological constant,” says cosmologist Jo Dunkley of Princeton University. “In 5 years, I hope we’ll have firm evidence that lambda-CDM is broken in some way.” With luck, she may also have a glimmer of what might take its place.