Could Smooth Cosmic Filaments Help Distinguish Warm Dark Matter?

At first glance, the space between galaxies looks empty. It is not. Threads of gas and dark matter—cosmic filaments—funnel matter into galaxies and shape the largest structures in the universe. Dark matter itself emits no light, so astronomers infer it from how it bends starlight and pulls gas into place. A new arXiv preprint asks whether those filaments might carry a more direct clue to what dark matter actually is.

The standard model, ΛCDM, treats dark matter as cold: particles so slow that they clump into structures as small as Earth-mass halos. That model has explained the cosmic microwave background and the large-scale distribution of galaxies. But on small scales, it faces persistent problems, including the missing satellites, cusp-core, and too-big-to-fail challenges. Ordinary-matter processes such as cosmic reionization and stellar feedback might alleviate some of these. Warm dark matter offers another route: particles with enough thermal motion to suppress small-scale structure.

The trouble is degeneracy. Complicated baryonic effects at galactic scales can mimic signatures of warm dark matter, so small-scale puzzles alone cannot settle the question. The paper, led by Yizhou Liu and colleagues and posted to arXiv on Jan. 30, 2026, looks for a probe less tangled in galactic physics: Lyman-alpha emission from cosmic filaments. Lyman-alpha, often shortened to Lyα, is ultraviolet light emitted when hydrogen atoms settle from the 2p excited state to the ground state. Because gas along the line of sight attenuates it, Lyα is a sensitive tracer of diffuse hydrogen in the intergalactic medium.

A Glow That Traces the Invisible

The question is simple to state: if dark matter is cold, do filaments glow differently than if it is warm? To find out, the team ran two high-resolution zoom-in hydrodynamical simulations of a Milky Way-like halo from the Aquarius project. The simulations were built from the same random phases, so initial fluctuations matched exactly. One used cold dark matter; the other used warm dark matter modeled as a 1.5 keV thermal relic particle. The team then produced Lyα surface-brightness maps at redshifts 2.5 and 4. In these simulations, filaments are denser at the higher redshift, z=4.

Redshift is a measure of cosmic time and distance: the higher the redshift, the farther back we look. The researchers computed Lyα emission from two processes: recombination, as ionized hydrogen and electrons combine, and collisional excitation, as free electrons strike neutral hydrogen. They corrected for self-shielding, where dense gas blocks ionizing radiation, and applied line-of-sight transmission factors of 0.92 at redshift 2.5 and 0.52 at redshift 4.

At Redshift 4, the Difference Is Sharpest

The results are a forecast, not an observation. In the simulations, warm dark matter filaments are smoother and denser in neutral hydrogen, and smoother and brighter in Lyα, than cold dark matter filaments. The contrast is strongest at redshift 4. By redshift 2.5, small Lyα clumps begin to appear along the warm dark matter filament, reducing the difference. That matters: it suggests high-redshift filaments are the better hunting ground for distinguishing dark matter models.

The authors also asked what current and future telescopes could see. They added noise to mock Lyα maps at redshift 4 using sensitivities expected for MUSE on the Very Large Telescope and a MUSE-like instrument on the Extremely Large Telescope. MUSE is an integral field spectrograph, an instrument that takes a spectrum for each pixel across a small patch of sky. Warm dark matter filaments are marginally detectable with MUSE on the VLT at noise levels of about 3.5 × 10^-20 erg per second per square centimeter per arcsecond squared. Cold dark matter filaments are fainter and require higher sensitivity, achievable with an ELT-class 30-meter telescope at about 10^-20 in the same units.

Field of view—the patch of sky an instrument covers—matters even more than angular resolution. Faint, extended filaments need long exposures, and a wide view increases the chance of capturing long threads. MUSE's 1 by 1 arcminute field could cover filaments stretching 1.5 to 2 h⁻¹ comoving megaparsecs at these redshifts. That is one reason the authors argue that the upcoming generation of 30-meter-class telescopes will be critical.

Where the Forecast Could Bend

If future observations find smooth, extended Lyα filaments, the result could favor warm dark matter or fuzzy dark matter, which suppresses small-scale structure similarly. If they find clumpy filaments, that would fit cold dark matter. But the study is careful about its assumptions. The simulations do not include baryonic feedback, because star formation efficiency inside filaments and feedback in warm dark matter environments are poorly constrained. The Lyα emission model is intentionally conservative, yielding a lower bound on surface brightness. The warm dark matter particle mass of 1.5 keV is lower than current Lyman-alpha forest constraints, chosen to make differences visible. And existing detections of diffuse Lyα filaments, such as those in the MUSE Ultra Deep Field, are rare, so they cannot yet settle the question.

Those caveats do not make the idea empty. The authors argue that even if feedback changes the neutral hydrogen in filaments, it should act differently in warm and cold dark matter because their small-scale structure differs. The relative contrast between models should remain a useful discriminant.

This work suggests that the cosmic web's filaments may have a faint Lyα glow—one that could help reveal whether its invisible scaffolding is built from cold, slow particles or warmer, faster ones. Seeing the clearest contrast, especially the fainter cold-dark-matter filaments, will likely take the next generation of 30-meter-class telescopes, and the patience to stare at nearly empty space until the filaments appear.