News by Date
News by Category
Solar System
White Dwarfs
Neutron Stars
Black Holes
Milky Way Galaxy
Normal Galaxies
Galaxy Clusters
Cosmology/Deep Field
Press Resources
Status Reports
Press Advisories
Image Releases
Release Guidelines
Image Use Policy
Web Shortcuts
Chandra Blog
RSS Feed
Email Newsletter
News & Noteworthy
Image Use Policy
Questions & Answers
Glossary of Terms
Download Guide
Get Adobe Reader
Related Links

Chandra @ NASA
Visit the Chandra pages at the NASA portal (opens in new window)
Image Use
Image Use Policy & Request Form
Guidelines for utilizing images, applets, movies, and animations featured in this Web Site.
Universe's Expansion May Not Be The Same In All Directions

For Release: April 8, 2020


Credit: NASA/CXC/Univ. of Bonn/K. Migkas et al.
Press Image, Caption, and Videos

(April 9th, 2020: Please see the end of this release for an added note discussing the possibility of systematic errors.)

One of the fundamental ideas of cosmology is that everything looks the same in all directions if you look over large enough distances. A new study using data from NASA's Chandra X-ray Observatory and ESA's XMM-Newton is challenging that basic notion.

Astronomers using X-ray data from these orbiting observatories studied hundreds of galaxy clusters, the largest structures in the universe held together by gravity, and how their apparent properties differ across the sky.

"One of the pillars of cosmology — the study of the history and fate of the entire universe — is that the universe is 'isotropic,' meaning the same in all directions," said Konstantinos Migkas of the University of Bonn in Germany, who led the new study. "Our work shows there may be cracks in that pillar."

Astronomers generally agree that after the Big Bang, the cosmos has continuously expanded. A commonly analogy is that this expansion is like a baking loaf of raisin bread. As the bread bakes, the raisins (which represent cosmic objects like galaxies and galaxy clusters) all move away from one another as the entire loaf (representing space) expands. With an even mix the expansion should be uniform in all directions, as it should be with an isotropic universe. But these new results may not fit that picture.

"Based on our cluster observations we may have found differences in how fast the universe is expanding depending on which way we looked," said co-author Gerrit Schellenberger of the Center for Astrophysics | Harvard & Smithsonian (CfA) in Cambridge, Massachusetts. "This would contradict one of the most basic underlying assumptions we use in cosmology today."

Scientists have previously conducted many tests of whether the universe is the same in all directions. These included using optical observations of exploded stars and infrared studies of galaxies. Some of these previous efforts have produced possible evidence that the universe is not isotropic, and some have not.

This latest test uses a powerful, novel and independent technique. It capitalizes on the relationship between the temperature of the hot gas pervading a galaxy cluster and the amount of X-rays it produces, known as the cluster's X-ray luminosity. The higher the temperature of the gas in a cluster, the higher the X-ray luminosity is. Once the temperature of the cluster gas is measured, the X-ray luminosity can be estimated. This method is independent of cosmological quantities, including the expansion speed of the universe.

Once they estimated the X-ray luminosities of their clusters using this technique, scientists then calculated luminosities using a different method that does depend on cosmological quantities, including the universe's expansion speed. The results gave the researchers apparent expansion speeds across the whole sky — revealing that the universe appears to be moving away from us faster in some directions than others.

The team also compared this work with studies from other groups that have found indications of a lack of isotropy using different techniques. They found good agreement on the direction of the lowest expansion rate.

The authors of this new study came up with two possible explanations for their results that involve cosmology. One of these explanations is that large groups of galaxy clusters might be moving together, but not because of cosmic expansion. For example, it is possible some nearby clusters are being pulled in the same direction by the gravity of groups of other galaxy clusters. If the motion is rapid enough it could lead to errors in estimating the luminosities of the clusters.

These sorts of correlated motions would give the appearance of different expansion rates in different directions. Astronomers have seen similar effects with relatively nearby galaxies, at distances typically less than 850 million light years, where mutual gravitational attraction is known to control the motion of objects. However, scientists expected the expansion of the universe to dominate the motion of clusters across larger distances, up to the 5 billion light years probed in this new study.

A second possible explanation is that the universe is not actually the same in all directions. One intriguing reason could be that dark energy — the mysterious force that seems to be driving acceleration of the expansion of the universe — is itself not uniform. In other words, the X-rays may reveal that dark energy is stronger in some parts of the universe than others, causing different expansion rates.

"This would be like if the yeast in the bread isn't evenly mixed, causing it to expand faster in some places than in others," said co-author Thomas Reiprich, also of the University of Bonn. "It would be remarkable if dark energy were found to have different strengths in different parts of the universe. However, much more evidence would be needed to rule out other explanations and make a convincing case."

Either of these two cosmological explanations would have significant consequences. Many studies in cosmology, including X-ray studies of galaxy clusters, assume that the universe is isotropic and that correlated motions are negligible compared to the cosmic expansion at the distances probed here.

The team used a sample of 313 galaxy clusters for their analysis, containing 237 clusters observed by Chandra with a total of 191 days of exposure, and 76 observed by XMM-Newton, with a total of 35 days of exposure. They also combined their sample of galaxy clusters with two other large X-ray samples, using data from XMM-Newton and the Japan-US Advanced Satellite for Cosmology and Astrophysics (ASCA), giving a total of 842 different galaxy clusters. They found a similar result using the same technique.

A paper describing these results will appear in the April 2020 issue of the journal Astronomy and Astrophysics and is available online. In addition to Migkas, Schellenberger and Reiprich, the authors of this paper are Florian Pacaud and Miriam Elizabeth Ramos-Ceja (University of Bonn), and Lorenzo Lovisari (CfA).

NASA's Marshall Space Flight Center manages the Chandra program. The Smithsonian Astrophysical Observatory's Chandra X-ray Center controls science and flight operations from Cambridge and Burlington, Massachusetts.

Other materials about the findings are available at:

For more Chandra images, multimedia and related materials, visit:

Media contacts:
Megan Watzke
Chandra X-ray Center, Cambridge, Mass.

Added note (April 9th, 2020): In the original posting of this press release we did not mention the possibility of "systematic errors" explaining some or all of these results, rather than the two cosmological explanations we discuss. First author Konstantinos Migkas had explained an example of a possible systematic error in his blog post that accompanied the press release. A relevant excerpt is here:

"So did we tear down one of the most crucial pillars of cosmology? Not so fast, it is not that simple. At least two scenarios may have led us to wrong conclusions.

Firstly, cosmic material might interfere with the light that travels from the clusters to the Earth. For example, previously unknown gas and dust clouds beyond the Milky Way could obscure a fraction of photons emitted from the clusters. Since we ignore the possible existence of such clouds, we do not account for their interference, and hence we would falsely underestimate the true luminosity of the clusters. Eventually, we could mistake this for a cosmological effect. We performed several tests that led us to believe that this scenario seems unlikely, but not impossible. However, considering that the direction of the anisotropy we find agrees with other studies that used observations in light at different wavelengths, where such obscuring effects are not expected, one could argue against the possibility of such biases in our analysis."

Their peer reviewed and published paper considers this and several other possible systematic errors in detail, and will address other possibilities in a future peer-reviewed paper.