The cosmic web resembles a giant three-dimensional network: galaxies and clusters are connected by filaments of dark matter. These filaments act as invisible bridges along which matter flows toward major nodes. For a long time, scientists struggled to understand the density structure of these filaments, but results were contradictory.
In the new study, the authors used data from the TNG50 simulation, which tracks the evolution of the universe, and extracted filaments using the DisPerSE algorithm. It turned out that directly identified filament centers do not coincide with the actual density peaks. The researchers then applied a "shrinking cylinder" method: they iteratively shifted the cylinder axis toward the density maximum until they found the true filament spine. This allowed them to recover the inner power-law slope of the density profile.
Next, the authors split dark matter particles into a "smooth" component and particles bound in small halos. It turned out that the central density peak in filaments is created by these small halos, while the smooth component has a flat core within 0.1 virial radius. This means that clumps dominate the filament centers, not a uniform distribution.
The study also showed that over time, the smooth component becomes less pronounced: in the early universe, matter accretion along filaments was smoother, whereas by now more matter falls in as clumps. This transition from smooth flow to clumpy accretion reflects the overall evolution of cosmic structure.
In the end, the authors proposed a generalized model with three power laws that fits the universal filament density profile well. This finding highlights that the cosmic web, like dark matter halos, obeys the principle of self-similarity: you just need to find the right scale.
If you could see the filaments of the cosmic web, they would resemble a glowing neural network stretched over billions of light-years, but made mostly of invisible dark matter.