getting to the "point of no return" Hints grow stronger that dark energy changes over time Latest DESI results fall short of discovery threshold but strengthen evidence for dynamical dark energy. Jennifer Ouellette Mar 19, 2025 6:00 pm | 1 DESI maps distant objects to study dark energy. Credit: KPNO/NOIRLab/NSF/AURA/B. Tafreshi DESI maps distant objects to study dark energy. Credit: KPNO/NOIRLab/NSF/AURA/B. Tafreshi Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreLast year, we reported on an exciting hint of new physics in the first data analysis results from the Dark Energy Spectroscopic Instrument (DESI)namely that the dark energy, rather than being constant, might vary over time. Granted, those hints were still below the necessary threshold to claim discovery and hence fell under the rubric of "huge, if true."But now we have more data from DESI, combined with other datasets, and those hints have gotten significantly strongerso much so that Mustapha Ishak-Boushaki of the University of Texas at Dallas, who co-chairs one of the DESI working groups, said that "we are getting to the point of no return" for confirming dynamical dark energy. Ishak-Boushaki and several other DESI team members presented their results at the American Physical Society's Global Physics Summit today in Anaheim, California. Several relevant papers have also been posted to the physics arXiv.Einsteins cosmological constant (lambda) implied the existence of a repulsive form of gravity. (For a more in-depth discussion of the history of the cosmological constant and its significance for dark energy, see our 2024 story.) Quantum physics holds that even the emptiest vacuum is teeming with energy in the form of virtual particles that wink in and out of existence, flying apart and coming together in an intricate quantum dance. This roiling sea of virtual particles could give rise to dark energy, giving the Universe a little extra push so that it can continue accelerating. The problem is that the quantum vacuum contains too much energy: roughly 10120 times too much.So the Universe should be accelerating much faster than it is if the dark energy is, essentially, the cosmological constant. Still, all the observations to date indicate that it's constant. The best theoretical fit thus far is known as the Lambda CDM model, which incorporates both a weakly interacting cold dark matter and dark energy. One alternative theory proposes that the Universe may be filled with a fluctuating form of dark energy dubbed quintessence. There are also several other alternative models that assume the density of dark energy has varied over the history of the Universe.In its earliest days, the Universe was a hot, dense soup of subatomic particles, including hydrogen and helium nuclei, aka baryons. Tiny fluctuations created a rippling pattern through that early ionized plasma, which froze into a three-dimensional place as the Universe expanded and cooled. Those ripples, or bubbles, are known as baryon acoustic oscillations (BAO). It's possible to use BAOs as a kind of cosmic ruler to investigate the effects of dark energy over the history of the Universe. DESI is a state-of-the-art instrument that can capture light from up to 5,000 celestial objects simultaneously. That's what DESI was designed to do: take precise measurements of the apparent size of these bubbles (both near and far) by determining the distances to galaxies and quasars over 11 billion years. That data can then be sliced into chunks to determine how fast the Universe was expanding at each point of time in the past, the better to model how dark energy was affecting that expansion.An upward trendLast year's results were based on analysis of a full year's worth of data taken from seven different slices of cosmic time and include 450,000 quasars, the largest ever collected, with a record-setting precision of the most distant epoch (between 8 to 11 billion years back) of 0.82 percent. While there was basic agreement with the Lamba CDM model, when those first-year results were combined with data from other studies (involving the cosmic microwave background radiation and Type Ia supernovae), some subtle differences cropped up.Essentially, those differences suggested that the dark energy might be getting weaker. In terms of confidence, the results amounted to a 2.6-sigma level for the DESI's data combined with CMB datasets. When adding the supernovae data, those numbers grew to 2.5-sigma, 3.5-sigma, or 3.9-sigma levels, depending on which particular supernova dataset was used.It's important to combine the DESI data with other independent measurements because "we want consistency," said DESI co-spokesperson Will Percival of the University of Waterloo. "All of the different experiments should give us the same answer to how much matter there is in the Universe at present day, how fast the Universe is expanding. It's no good if all the experiments agree with the Lambda-CDM model, but then give you different parameters. That just doesn't work. Just saying it's consistent to the Lambda-CDM, that's not enough in itself. It has to be consistent with Lambda-CDM and give you the same parameters for the basic properties of that model."These latest results cover the first three years of collected data, spanning almost 15 million galaxies and quasars. Once again, the DESI data alone was consistent with Lambda CDM, i.e., the dark energy is constant. And once again, when combined with other datasetsfrom CMB, supernovae, and weak gravitational lensing studiesstrong hints emerged that dark energy might be changing over time. The confidence level ranges from 2.8 to 4.2 sigma, depending on the combination of datasetsjust shy of the five-sigma threshold.This might strike the average citizen as an incremental advance, but the reality is more complicated. "The DESI data itself is not incremental" said Percival. "We now have three years of data rather than one year of data. That is substantial, not just because of an increased area but because we've increased the overlap. The way we do the survey is we build up plates on the sky, and, after three years rather than one year of operations, we have a lot more of those overlaps filled in. So our data is a lot more complete in the sense that we've gone down to the full depth that we expect to get to in more patches. Consequently, our BAO measurements themselves are a lot better. They're between a factor of two and three better depending on exactly this balance between area versus depth." A slice of the DESI data mapping celestial objects from Earth (center) to billions of light years away. Credit: Claire Lamman/DESI Collaboration Catherine Heymans, Astronomer Royal of Scotland, told Ars that these new results give scientists much more confidence in DESI's analysis. She was surprised at the excitement over last year's first results, since, "whenever there's a first data release, the scientific community always takes the results with a pinch of salt," she said. But DESI made their data public, and other scientists have been making their own analyses over the last year; it has stood up to that close scrutiny."The really strong significance for dynamical dark energy comes from the combination of the DESI standard ruler, the BAO plus the supernova data," she added. "That's two different ways of measuring the expansion rate of the Universe. By combining those two things together, you get this strong detection of dynamical dark energy."The next step for the DESI collaboration is to analyze five years' worth of data to see if the upward trend toward the 5-sigma threshold for discovery holdsperhaps even surpassing that threshold, which would be very exciting indeed. That will likely not happen for another two years, per Percival. Should 5 sigma be reached, Heymans said astronomers should expect to see similar results in data from the Euclid Space Telescope, which is slated to do a similar experiment to DESI, at higher redshift, in the near future."It opens up a huge range of possibilities," said Percival of the implications should it be confirmed that dark energy changes over time. "It will keep theorists happy for many years to come. As a scientist you want to sit a little bit on the fence. But if this is right, this is the next step after the discovery of dark energy. Lambda works. Now, Lambda doesn't work. It means there's a lot more information that's accessible about this process. I think people were worried that everything would show that it just exactly agrees with Lambda. But if there's actually things happening to how the acceleration is changing within detail, that's exciting because we can get a handle on the physics.""There's no fundamental underpinning for what could be causing that dynamical dark energy and that does make me anxious," said Heymans. "It's like the observers are throwing the gauntlet back to the theorists. It'd be nice to be able to explain two dark entities with one fell swoop. I am excited about cracks in the cosmological model because this way is pushing the theoretical community to think outside the box to think of new ideas. And maybe that will solve the whole dark entity conundrum, which is why we're all here."Jennifer OuelletteSenior WriterJennifer OuelletteSenior Writer Jennifer is a senior writer at Ars Technica with a particular focus on where science meets culture, covering everything from physics and related interdisciplinary topics to her favorite films and TV series. Jennifer lives in Baltimore with her spouse, physicist Sean M. Carroll, and their two cats, Ariel and Caliban. 1 Comments