Duke Scientists Help Sharpen the Picture of a Changing Universe

An international team of researchers with the Dark Energy Survey (DES) has released one of the most precise new measurements yet of how the universe expands over time. The results, based on hundreds of millions of galaxies, help scientists test the leading model of cosmology and probe the mysterious force known as dark energy that appears to be driving cosmic acceleration.

Among the scientists shaping this latest analysis are Duke researchers, led by Associate Professor of Physics Michael Troxel, whose group has played a central role in turning faint traces of light into a high-precision map of the invisible universe.

Measuring the invisible

Roughly 95% of the universe is made of dark matter and dark energy — components that cannot be seen directly. Instead, astronomers must infer their presence through subtle effects on visible matter and light.

Troxel’s research focuses on a technique called weak gravitational lensing, also known as “cosmic shear.” As light from distant galaxies travels toward Earth, it passes through the gravitational pull of intervening matter. That mass — mostly dark matter — slightly bends and stretches the light, distorting the shapes of background galaxies by tiny amounts.

“These distortions are incredibly small,” Troxel explained in a previous interview. “But when you measure them for hundreds of millions of galaxies, you can reconstruct how matter is distributed across the universe.”

The new Dark Energy Survey results combine weak lensing with other probes (methods used to estimate distances across the cosmos), including supernova measurements and galaxy clustering, to test whether today’s standard cosmological model continues to hold up under increasingly precise data.

Duke’s imprint on the data 

Several members of Duke’s Cosmology group have helped build the essential tools that make possible such precise measurements of the cosmos.

Masaya Yamamoto, a former Duke Ph.D. student now a postdoctoral researcher at Princeton University, co-led the development of the survey’s state-of-the-art shape catalog while at Duke. This catalog quantifies the minute distortions in galaxy shapes caused by weak lensing — the foundational dataset for cosmic shear studies. Yamamoto also co-led a final cosmic shear cosmology paper, presenting the collaboration’s most comprehensive lensing-only cosmological constraints.

“In this new analysis — based on 150 million galaxies with well-measured shapes — we observe galaxies further back in time and see more and more structure between those distant galaxies and us,” said Yamamoto. “This led to an almost twofold increase in the precision with which we understand how structure grows in our universe.”

Producing a reliable shape catalog is not simply a matter of measuring galaxy outlines. Telescopes blur images, atmospheric turbulence introduces distortions and detectors add noise. Yamamoto and collaborators developed sophisticated calibration methods to separate genuine cosmic signals from these confounding effects, greatly increasing the precision of weak lensing measurements.

Because astronomers cannot measure precise distances to hundreds of millions of galaxies individually, they estimate those distances from galaxy colors — a technique known as “photometric redshifts.” Boyan Yin, a current Duke Ph.D. student, co-led the development of improved photometric redshift methods used in the new analysis. By refining how galaxy colors translate into distance estimates, Yin’s work reduces one of the largest sources of uncertainty in weak lensing studies. Better distance estimates mean sharper maps of matter, and more reliable conclusions about the universe’s expansion history.

“It is fascinating that from the Dark Energy Survey images of different colors — which are not that different from the images you take with a modern phone — we can probe the redshift of hundreds of millions of galaxies, and thus infer the history of the universe,” said Yin.

Supernovae and cosmic acceleration

The Duke impact also extends beyond lensing. Brodie Popovic, a former Duke Ph.D. student now a postdoctoral researcher at the University of Southampton, led a final re-analysis of the survey’s supernova results. Type Ia supernovae — exploding stars with a very consistent brightness — serve as “standard candles” for measuring cosmic distances. By comparing how bright they appear to how bright they should be, astronomers can determine how fast the universe was expanding at different epochs. Popovic’s work adds precision to the supernova-based results, allowing them to be combined consistently with lensing and galaxy clustering measurements.

When multiple independent methods point to the same cosmological parameters, confidence in the results grows. When they disagree, physicists take notice. The new analysis helps clarify where current measurements align — and where subtle tensions with other surveys may still remain.

Testing dark energy — and gravity itself

Sujeong Lee, a former Duke postdoctoral researcher now a professor at Ohio University, is co-leading a forthcoming study that uses the same dataset to probe evolving dark energy and possible modifications to gravity. Early results were presented at a recent meeting of the American Astronomical Society.

If dark energy changes over time — or if gravity behaves differently on the largest cosmic scales — those deviations could appear in the combined lensing and supernova data. Such tests push beyond measuring cosmological parameters, and ask whether our current understanding of physics holds at the largest scales imaginable.

A training ground for the next generation

For Troxel, who is also the co-director of the SPACE Initiative at Duke, the significance of these results lies not only in the cosmology, but also in the collaborative training environment that made them possible. Large surveys like the Dark Energy Survey are massive international efforts, requiring experts in instrumentation, data processing, statistical analysis and theoretical modeling. Duke students and postdocs have led significant advances in these areas and made vital contributions to the reliability of the final results.

As even larger observatories come online in the coming years, the methods pioneered in this work will serve as the foundation for even more precise tests of how the universe expands. Arun Kannawadi, assistant research professor of Physics and part of the team working on the Rubin Observatory, points to that connection. “Many of the algorithms developed by the DES team have been adopted for the weak lensing analyses of the first year of Rubin’s Legacy Survey of Space and Time (LSST), because they have proven to be the best so far,” he said. 

“The volume of data from Rubin Observatory will eventually dwarf that from DES, but DES’s true legacies are its methodology development and the people it trained to carry out the next generation of survey science.”