Messier 79 globular star cluster, Hubble image

Ageing red-giant stars are the focus of a new method that measures the Universe’s current rate of expansion.Credit: NASA/ESA/SPL

For much of this decade, the two most precise gauges of the Universe’s rate of expansion have been in glaring disagreement. Now, a highly anticipated independent technique that cosmologists hoped would solve the conundrum is instead adding to the confusion.

In results unveiled1 on 16 July and due to appear in the Astrophysical Journal, a team led by astronomer Wendy Freedman at the University of Chicago in Illinois presents a technique that measures the expansion using red-giant stars. It had promised to replace a method that astronomers have been using for more than a century — but for now, the speed measurement has failed to resolve the dispute because it falls half way between the two contentious values.

“The Universe is just messing with us at this point, right?” tweeted one astrophysicist about the paper.

“Right now, we’re trying to understand how it all fits together,” Freedman told Nature. If the cosmic-speed discrepancy is not resolved, some of the basic theories that cosmologists use to interpret their data — such as assumptions about the nature of dark matter — could be wrong. “Fundamental physics hangs in the balance,” Freedman says.

Cosmic speedometer

American astronomer Edwin Hubble and others discovered in the 1920s that the Universe is expanding by showing that most galaxies are receding from the Milky Way — and the farther away they are, the faster they are receding. The roughly constant ratio between speed and distance became known as the Hubble constant. For each additional megaparsec (around 3.26 million light years) of distance, Hubble found that galaxies receded 500 kilometres per second faster — so the Hubble constant was 500 in units of kilometres per second per megaparsec.

Over the decades, astronomers substantially revised down the estimate as measurement techniques improved. Freedman pioneered the use of the Hubble Space Telescope in the 1990s to (fittingly) measure the Hubble constant, and calculated a value of around 72 with an error margin of around 10%. A team led by Nobel laureate Adam Riess at Johns Hopkins University in Baltimore, Maryland, has made the most precise measurements so far, and its latest value is 74, with an error margin of just 1.91%2.

But a separate effort in the past decade has thrown a spanner in the works. Scientists with the European Space Agency’s Planck mission mapped the relic radiation of the Big Bang, called the cosmic microwave background, and used it to calculate the Universe’s basic properties. Using standard theoretical assumptions about the cosmos, they calculated the Hubble constant as 67.8.

The difference between 67.8 and 74 might seem small, but it has become statistically significant as both techniques have improved. So, theorists have started to wonder whether the reason for the discrepancy lies in the standard theory of cosmology, called ΛCDM, which assumes the presence of invisible particles of dark matter as well as a mysterious repulsive force called dark energy. But they have struggled to find a tweak to the theory that could solve the problem and still be consistent with everything that is known about the Universe. “It’s hard to look at ΛCDM and see where the loose threads are, that if you pull them, they will unravel it,” says Rocky Kolb, a cosmologist at the University of Chicago.

Freedman’s technique updates a key element of the established Hubble measurement method — and produces a value of 69.8.

The hard part of measuring the Hubble constant is to reliably gauge galaxies’ distances. Hubble’s first estimate depended on measuring the distances of nearby galaxies by observing individual, bright stars called Cepheids. Astronomer Henrietta Swan Leavitt had discovered in the early twentieth century that these stars’ actual brightness was predictable. So, by measuring how bright they appeared on photographic plates, she could calculate how far away the stars were. Astronomers call such signposts standard candles.

But researchers have since been trying to find better standard candles than Cepheids, which tend to exist in crowded, dust-filled regions that can distort estimates of their brightness. “The only way we have to get to the bottom on this is to have independent methods, and up to this point we’ve had no checks on the Cepheids,” says Freedman, who has spent much of her career improving the precision and accuracy of Cepheid measurements. “She knows where all the bodies are buried,” says Kolb.

Freedman and her colleagues sidestepped Cepheids altogether, and instead used as their standard candles red giants — old stars that have become puffed out — together with supernovae explosions, which serve as signposts for more-distant galaxies.

Giant calculation

Red giants are more common than Cepheids, and are easy to spot in the peripheral regions of galaxies, where stars are well separated from one another and dust is not an issue. Their brightness varies widely — but, taken as a whole, a galaxy’s red-giant population has a handy feature. The stars’ brightness increases over millions of years until it reaches a maximum, and then it suddenly drops. When astronomers plot a large group of stars by colour and brightness, the red giants look like a cloud of dots with a sharp edge. The stars at that edge can then serve as standard candles.

Freedman’s team used the technique to calculate the distances to 18 galaxies, and obtained an estimate of the Hubble constant that for the first time has an accuracy comparable with that of the Cepheid-based studies.

Riess says that the red-giant study still relies on assumptions about the amount of dust in galaxies — particularly in the Large Magellanic Cloud, which the study used as an anchor point. “Dust is very tricky to estimate, and I am sure there will be lots of discussion” about why the authors’ approach leads to a lower estimate of the Hubble constant, he says.

The result is statistically compatible with the Planck prediction and with Riess’s Cepheid calculation — meaning that the error bars of the calculations overlap — and the technique’s precision will improve as data on red giants accumulate. They could beat Cepheids in the near future, Kolb says.

The needle could shift towards one of the other values. Or it could stay put, and the other techniques might eventually converge to it. For now, cosmologists have plenty to puzzle over.