Dark matter may consist of particles with different masses

Dark matter keeps getting blamed for the universe’s big patterns while staying stubbornly out of reach. You cannot see it, touch it, or capture it. Yet its gravity helps shape galaxies, and that makes it hard to ignore.

For decades, many astronomers have leaned on the cold dark matter model to explain how cosmic structure forms. The model works well on large scales. Trouble starts when you zoom in. As measurements sharpen, a set of small-scale puzzles has piled up that does not sit neatly inside the classic picture.

Some dwarf galaxies, for example, seem to have dark matter that spreads out, with surprisingly low density in the center. At the same time, strong gravitational lensing has uncovered dark substructures that look almost too compact, far denser than traditional expectations. These two signals have lived side by side for years, and they pull in opposite directions.

A team of physicists from Purple Mountain Observatory, CAS, offers a way to hold both ideas at once. Their new work argues that dark matter may not be a single ingredient. It may include at least two kinds of particles with different masses.

The study focuses on what the authors call a “two-component self-interacting dark matter” model. In it, dark matter still moves through the universe under gravity, but the particles can also collide. That extra interaction matters because collisions between particles of different masses can sort them over time.

Heavier particles tend to drift inward, toward galaxy centers. Lighter particles more often end up farther out. The paper compares this to what happens in star clusters, where massive stars migrate inward while lower-mass stars occupy larger orbits.

That sorting process has a name in the study: mass segregation. It is not a small tweak. In the team’s framework, it becomes a unifying mechanism that can push dark matter halos in different directions depending on environment and history.

In the team’s simulations and theoretical modeling, mass segregation can create low-density cores in dwarf galaxies while still allowing some halos to become denser and more compact. That matters because the field has wrestled with how to explain “diffuse” dwarf galaxies and ultra-compact substructures without swapping in a different explanation for each one.

The paper places self-interacting dark matter, or SIDM, at the center of that effort. SIDM has long looked attractive because gravothermal evolution in these models can create either cored profiles or cuspy ones. It can also connect to a wide range of claimed oddities, from the diversity of rotation curves to very dense systems and even early supermassive black hole formation.

But the study’s authors stress a gap: a single, unified SIDM story has remained hard to pin down across all the small-scale phenomena people keep pointing at.

Part of the tension comes from observations that tug in different directions. Dwarf galaxy clustering suggests a population with growing cores. Strong lensing reconstructions of compact substructures in systems such as SDSS J0946+1006 and JVAS B1938+666 suggest core-collapsed dwarf halos instead. And galaxy-galaxy strong lensing measurements reported by Meneghetti and colleagues have indicated too many small-scale lensing events compared with cold dark matter expectations, by a factor of 3 to 6.

Not everything that boosts lensing helps in the right way, either. The study notes that SIDM halos with baryons have been found to mainly enhance two-image lensing events, rather than the four-image events observed.

Here is where the two-component idea does extra work. In the two-component case, collisions between the heavy and light species transfer energy from the heavier to the lighter species. The study argues that this can raise inner densities while still producing cores in dwarf halos.

And unlike one-component SIDM, the mass enclosed within the Einstein radius can grow in this picture. That detail matters for lensing, because it can strengthen the lensing signal that dense substructures create.

The paper describes several modeling choices and simulation setups to explore this. The authors simulate halos under different two-component SIDM variants, including a velocity-dependent model and an extreme case that keeps only inter-species interactions. They also build a parametric model to extend predictions beyond the resolution limit of the simulations, and they take a conservative approach in lensing inferences.

To set up initial conditions without changing the system’s total mass, the simulations randomly assign half the particles to be 50% more massive and half to be 50% less massive. A collisionless stellar component modeled by a Hernquist profile appears in some runs meant to represent baryonic effects.

The results tie back to the observational tensions. The study reports that mass segregation can help form dense, lensing-relevant configurations at modest cross sections, where one-component models would struggle to reach the same compactness. In their strong lensing analysis, the two-component models increase the number and size of small-scale caustics compared to cold dark matter.

Quantitatively, the paper compares a fiducial GGSL cross section across several scenarios. In the table they report, cold dark matter has a mean GGSL cross section of 0.3024. Model-2v rises to 0.6133, about 2.03 times the cold dark matter value. Model-x rises to 2.223, about 7.35 times. When baryons are included via their parametric approach, Model-2v+baryons reaches 1.360, about 4.50 times, and Model-x+baryons reaches 4.149, about 13.7 times.

This is not framed as a final victory lap. The authors spell out limits. Simulation resolution still constrains what can be said about dwarf halos. Their lensing analysis relies on a single cluster zoom evaluated across 11 projections. They also describe uncertainties tied to host-dependent properties, and they say more dedicated simulations are needed to map baryonic effects in full.

Still, the paper’s central bet is clear: those apparently clashing small-scale signals may point to one underlying theme. Dark matter’s internal properties may be more complex than a single cold, collisionless component.

This study was recently published in Science Bulletin. It is the Purple Mountain Observatory team’s second work on two-component self-interacting dark matter. An earlier study on mass segregation and the diversity of core density distributions in dwarf galaxies was published in Physical Review D. The authors listed here include Daneng Yang, Yi-Zhong Fan, Siyuan Hou, and Yue-Lin Sming Tsai.

If the universe’s missing matter has multiple components, the stakes extend beyond a tidy theory debate. Future surveys that map galaxies and lensing signals at higher precision could use “cosmic magnifying glasses” to test whether dark matter really sorts itself by mass. Better tests could narrow the range of viable particle models, and guide what kinds of interactions physicists should look for.

On the astronomy side, a unified explanation could change how researchers interpret small, stubborn mismatches between simulations and observations.

Instead of treating each anomaly as a separate problem, teams could use one framework to predict where diffuse cores should appear, and where dense substructures should concentrate.

Research findings are available online in the journal Science Bulletin.

The original story "Dark matter may consist of particles with different masses" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.