---
title: "Why did any dark matter survive the hot early universe?"
description: "Dark matter particles kept destroying each other until the expanding universe pulled them apart. How strongly they reacted set how many were left."
canonical_url: "https://sandboxphysics.com/stories/dark-matter-relics/why-some-dark-matter-survived-the-cooling/"
language: "en"
page_type: "story"
alternate_languages:
  zh-CN: "https://sandboxphysics.com/zh/stories/dark-matter-relics/why-some-dark-matter-survived-the-cooling/"
last_modified: "2026-11-14"
---

[How dark matter was left over](https://sandboxphysics.com/stories/dark-matter-relics/)

# Why did any dark matter survive the hot early universe?

Sandbox Physics · 2026-11-14

If dark matter is made of particles that once mixed with ordinary matter, they were made and destroyed in pairs while the young universe was hot. As it cooled, the destruction should have wiped them out. It did not, because space was expanding: the particles grew so far apart that pairs stopped meeting. The survivors were left over, and the strength of the reaction decided how many. A weaker reaction gives up sooner and leaves more behind.

## A hot soup that makes and destroys particles

About a billionth of a second after the start, the universe was a soup of particles at temperatures above a trillion kelvin. Collisions in such a soup are energetic enough to turn energy into pairs of heavy particles, and pairs of heavy particles can collide and turn back into energy. When both happen at the same rate, the number of heavy particles is said to be in equilibrium.

Equilibrium depends on temperature. While the soup is much hotter than the energy locked in a particle's mass, there are about as many heavy particles as photons. Once it cools below that energy, few collisions can still make a pair, yet existing pairs keep destroying each other. The equilibrium number then falls steeply, roughly halving for every few per cent drop in temperature.

If nothing else happened, a heavy dark particle would follow this falling number all the way down and almost none would remain today. Something else did happen.

## The race against expansion

The universe was also expanding, and very quickly. Two rates matter. One is how often a dark particle meets a partner and annihilates. The other is how fast space stretches. The first rate depends on how many partners are around, so it collapses as the equilibrium number falls. The second falls only gently as the universe cools.

Picture a dance floor that empties as the music fades, while the room itself keeps getting bigger. At first dancers find partners easily. Later the floor is so sparse and growing so fast that a dancer looking for a partner never finds one. The dancers left without partners are still there when the lights come on.

For a particle a hundred times heavier than a proton, with a reaction strength like that of the weak nuclear force, the reaction loses the race when the soup has cooled to about four billion electronvolts of thermal energy. In the teaching model on this site that happens about 17 billionths of a second after the start. After that the number per unit volume still falls, but only because space dilutes it, so the number in any expanding region stays fixed. Physicists call this freeze-out.

Fewer than one in a billion of the heavy particles that existed while the soup was hot survive it. That small remnant is enough. The Planck satellite measured how much dark matter there is today from the pattern of hot and cold spots in the oldest light, and a freeze-out of this kind leaves about that much. [Planck Collaboration, 2018, abstract](https://arxiv.org/abs/1807.06209v4)

## Why a weaker reaction leaves more

This is the surprise in freeze-out. A stronger reaction keeps the particles in equilibrium for longer, while their number is still falling, so it leaves fewer. Make the reaction a hundred times stronger and the leftover drops more than eightyfold. Make it weaker and the particles drop out sooner and more of them remain.

The reaction strength that leaves the observed amount is in the range people associate with particles near the weak-force scale. That is why weakly interacting massive particles, or WIMPs, became such popular candidates. The Particle Data Group review warns that the freeze-out formula itself has no built-in connection to the weak scale, despite the nickname "WIMP miracle". [Particle Data Group, 2025, section 27.3](https://pdg.lbl.gov/2025/reviews/rpp2025-rev-dark-matter.pdf)

## When the reaction never wins: freeze-in

Now turn the reaction down so far that it never keeps up with expansion, even when the soup is hottest. Two different stories are possible.

If the particles somehow started with their full equilibrium number, they drop out while they are still moving near light speed. Their number is then fixed at about the same as the number of photons. For a heavy particle that is far too much dark matter. In the Lab's model, only a particle lighter than about a hundred electronvolts could survive this way without overfilling the universe. Such light, fast particles make hot dark matter, which would smear out small galaxies and is ruled out. [Hall, Jedamzik, March-Russell and West, 2010, section 5.1](https://arxiv.org/abs/0911.1120v2)

If instead the particles started from nothing, the rare collisions in the hot soup make a few of them and almost none are destroyed. Production peaks when the soup cools to around the particle's mass energy and then stops. This is called freeze-in, and a stronger reaction now makes more particles, the opposite of freeze-out. [Hall, Jedamzik, March-Russell and West, 2010, section 1](https://arxiv.org/abs/0911.1120v2) In the Lab's model, a particle a hundred times heavier than a proton ends with the observed amount if its reaction is about eight billion billion times weaker than the freeze-out value.

So one equation gives three regimes as the reaction gets stronger. Freeze-in from nothing rises with strength. A hot plateau caps the amount. Freeze-out then falls as the strength grows. A strong enough reaction also erases the starting state: particles made from nothing reach equilibrium quickly and then freeze out exactly as if they had been there all along.

## If the early universe expanded faster

The expansion rate before the first atomic nuclei formed, about one second after the start, is not measured directly. Some theories add an extra form of energy that sped up the expansion at early times and then faded away.

Faster expansion makes the particles drop out of equilibrium earlier. One might expect that to leave more of them, and it does. But Francesco D'Eramo, Nicolas Fernandez and Stefano Profumo noticed something else. If the extra energy fades fast enough, the particles keep finding partners long after they first drop out, and their number keeps falling slowly until the extra energy is gone. They called this relentless dark matter. [D'Eramo, Fernandez and Profumo, 2017, abstract](https://arxiv.org/abs/1703.04793v1) In the Lab, the reaction strength that matches observations with normal expansion leaves about 34 times too much dark matter with such a phase. To match observations, the reaction would have to be much stronger.

## What the calculation does not tell us

The freeze-out picture assumes that dark matter is a particle, that it once mixed with ordinary matter, and that nothing else changed its number afterwards. None of these is established. Dark matter could be ultralight fields, primordial black holes or something not yet proposed. The calculation turns an assumption about early history into a prediction for one number, the amount of dark matter today. Many different stories can reproduce that one number.

The Lab simplifies more than a research code would. It holds the number of particle types in the soup fixed, while in reality it falls as the universe cools. For this particle that shifts the answer by less than ten per cent. It still shows the race clearly.

## Try it in the Lab

Open the [dark matter relic Lab](https://sandboxphysics.com/labs/dark-matter-freeze-out-freeze-in/) and watch the default run. Violet dots vanish in flashes while the cyan reaction bar towers over the orange expansion bar. When the cyan bar falls below the orange one, the flashes stop and the dots that remain stay. Drag the cyan handle up and run it again to see fewer survivors. Then choose Freeze-in from nothing and watch dots appear instead.

The next chapter, [Equilibrium, expansion and the reaction race](https://sandboxphysics.com/stories/dark-matter-relics/equilibrium-expansion-and-the-reaction-race/), puts numbers on each step.

## Sources

1. [Planck Collaboration, 2018, abstract](https://arxiv.org/abs/1807.06209v4)
2. [Particle Data Group, 2025, section 27.3](https://pdg.lbl.gov/2025/reviews/rpp2025-rev-dark-matter.pdf)
3. [Hall, Jedamzik, March-Russell and West, 2010, section 5.1](https://arxiv.org/abs/0911.1120v2)
4. [D'Eramo, Fernandez and Profumo, 2017, abstract](https://arxiv.org/abs/1703.04793v1)

## Keep reading

How dark matter was left over · Reading path

### From thermal equilibrium to relentless dark matter

1. [You are here **Why did any dark matter survive the hot early universe?**](https://sandboxphysics.com/stories/dark-matter-relics/why-some-dark-matter-survived-the-cooling/)
2. [Article **Equilibrium, expansion and the reaction race**](https://sandboxphysics.com/stories/dark-matter-relics/equilibrium-expansion-and-the-reaction-race/)
3. [Article **The Boltzmann equation and freeze-out**](https://sandboxphysics.com/stories/dark-matter-relics/the-boltzmann-equation-and-freeze-out/)
4. [Article **Freeze-in, hot relics and faster expansion**](https://sandboxphysics.com/stories/dark-matter-relics/freeze-in-hot-relics-and-faster-expansion/)
5. [Article **Recompute a relic abundance with the saved scripts**](https://sandboxphysics.com/stories/dark-matter-relics/recompute-a-relic-abundance/)
6. [Article **How to read the relentless dark matter paper of D'Eramo and colleagues**](https://sandboxphysics.com/stories/dark-matter-relics/reading-relentless-dark-matter/)
