Deep divefrom Episode 3Cosmology24 September 202614 min

Getting Warmer: IceCube Nears Freeze Out

A deep dive into one paper: what was known before, what the researchers did, what the numbers say and what they don't.

The paper: Getting Warmer — IceCube Nears Freeze Out (arXiv)

Chapters

Transcript

The script the two AI voices read from. The times are worked out from the script, not heard, so a line can start a few seconds early or late.

Read the whole episode

Cold open

Host 1Welcome to the Hastrup Review of Science — where artificial intelligence reads real science, so you don't have to pretend you understood the paper. One paper today. IceCube is a cubic-kilometre neutrino telescope buried in Antarctic ice. A new analysis uses its detection of neutrinos from our own galaxy to set some of the tightest limits yet on dark matter annihilation — and the authors argue that next-generation detectors could reach the critical threshold that would either confirm or rule out the most popular theory for how dark matter was produced in the early universe.

Host 2Mukhopadhyay and Krnjaic present two complementary methods for constraining dark matter annihilation into neutrinos using IceCube's twelve-year dataset. The paper is on the arXiv and has not yet been peer-reviewed.

What IceCube saw

Host 1The paper is called Getting Warmer: IceCube Nears Freeze Out. Before we get to the dark matter analysis, what did IceCube actually detect that made this possible?

Host 2In 2023, the IceCube collaboration announced with high statistical significance that it had detected a diffuse population of high-energy neutrinos coming from the plane of the Milky Way. This was a landmark result. Neutrinos had been detected from individual astrophysical sources before, but this was the first time a diffuse glow of neutrinos from the galaxy itself was confirmed.

Host 1What produces those neutrinos?

Host 2The conventional expectation is cosmic ray interactions. When high-energy cosmic rays — primarily protons and heavier nuclei — collide with interstellar gas in the galactic plane, they produce pions. Charged pions decay into muons and neutrinos. Neutral pions decay into gamma rays. So if you see diffuse gamma rays from the galactic plane, which we do through telescopes like Fermi-LAT, you expect a corresponding diffuse neutrino flux from the same processes.

Host 1And this paper asks whether some of those neutrinos could be coming from dark matter instead.

Dark matter and neutrinos

Host 2Exactly. If dark matter particles can annihilate — that is, if a dark matter particle encounters another and the two destroy each other, converting their mass into other particles — then one possible product is neutrinos. This paper focuses specifically on the channel where dark matter annihilates directly into neutrino-antineutrino pairs.

Host 1Why that channel specifically?

Host 2Because it is the hardest to probe. When dark matter annihilates into quarks or gauge bosons, the decay cascades produce gamma rays, and gamma-ray telescopes like Fermi-LAT have already placed very strong limits on those channels. But annihilation directly to neutrinos produces no gamma rays at all. The only way to detect it is with a neutrino telescope. Until IceCube confirmed this diffuse galactic neutrino signal, there was not enough data to competitively probe the neutrino channel using the galaxy as a whole.

Thermal freeze-out

Host 1The title says IceCube nears freeze-out. What is freeze-out?

Host 2Thermal freeze-out is the leading mechanism for how dark matter could have acquired its present abundance. In the very early universe, temperatures were so extreme that dark matter particles were constantly being created and annihilated in thermal equilibrium with all other particles. As the universe expanded and cooled, the temperature dropped below the dark matter mass, and creating new dark matter particles became energetically disfavoured. But annihilation continued until the dark matter density dropped so low that particles could no longer find each other efficiently. At that point, the abundance froze — it was locked in.

Host 1And we can calculate the annihilation rate that would produce the right amount of dark matter?

Host 2We can. The calculation gives approximately two times ten to the minus twenty-six cubic centimetres per second. This is called the thermal relic cross section. Any dark matter candidate that annihilates at roughly this rate would naturally produce the observed dark matter density — not too much, not too little. It is a remarkably clean prediction. And it is the specific number that this paper's analysis is approaching.

Host 1So detecting or excluding annihilation at that rate would be a major result.

Host 2It would. Despite decades of searching with neutrino telescopes, probing the thermal relic cross section for the neutrino annihilation channel has remained out of reach. Other channels have been constrained below the thermal benchmark using gamma rays, but the neutrino channel has lagged. What this paper demonstrates is that the detection of diffuse galactic neutrinos has suddenly made the thermal benchmark accessible.

The longitude test

Host 1Walk us through the two methods. The first one is called model-independent.

Host 2The key idea is spatial. Dark matter annihilation should produce more neutrinos from directions toward the galactic centre, where the dark matter density is highest, and fewer from directions away from it. Cosmic ray interactions also concentrate toward the centre, because that is where the gas is, but the two distributions have different shapes because the dark matter halo and the interstellar gas are distributed differently.

Host 1So they are looking at the pattern across the sky.

Host 2They divide the galactic plane into bins by galactic longitude and compare the observed neutrino excess in each bin — after subtracting the estimated atmospheric background — against what dark matter annihilation would predict. The critical virtue of this method is that it makes no assumption about which cosmic ray emission model is correct. It relies only on the spatial pattern.

Host 1What data are they using?

Host 2Twelve years of IceCube shower events with reconstructed energies above five TeV. In the innermost longitude bin — within twenty degrees of the galactic centre — they see two hundred and seventeen events against an atmospheric background of about one hundred and fifty-four, yielding an excess of roughly sixty-three events. This excess is what they partition between cosmic ray and potential dark matter contributions.

Host 1And the limits from this method?

Host 2For dark matter masses between one and three hundred TeV, this approach improves on previous IceCube limits by factors of several. It is a robust, conservative result. But on its own it does not quite reach the thermal relic cross section.

The template method

Host 1The second method does get closer.

Host 2The template-dependent method is more aggressive. Instead of using only the spatial distribution, it uses the full neutrino energy spectrum from the inner galaxy. The approach is to take the best-fit neutrino spectrum predicted by a given cosmic ray model and ask, at each energy, how much room remains for an additional dark matter signal on top of it.

Host 1But that requires choosing a cosmic ray model.

Host 2It does, and that is both its power and its limitation. They test four different models. The Fermi-LAT pi-zero template assumes spatially uniform cosmic ray diffusion and a spectrum falling as roughly E to the minus two point seven. The two KRA-gamma models use spatially dependent cosmic ray transport with different maximum rigidities, which changes how hard the spectrum extends. And the CRINGE model, a global cosmic ray fit that includes contributions from unresolved sources.

Host 1And the result?

Host 2Under several of these templates, the remaining room for a dark matter contribution approaches or touches the thermal relic cross section at certain masses in the TeV range. The exact mass range depends on the template, but the paper shows that the measured inner-galaxy neutrino intensity is already comparable to what you would expect from thermal-relic dark matter annihilation over part of this mass range.

Host 1That is an order-of-magnitude improvement over previous constraints?

Host 2In parts of the mass range, yes. The improvement is driven almost entirely by the existence of the diffuse galactic neutrino signal. Previous limits relied on looking for dark matter on top of an expected background with no confirmed galactic neutrino detection. Now there is a measured signal, and you can ask how much dark matter can hide inside it.

A word from our sponsor

Host 1Speaking of cubes and ice. Have you ever considered the kind that fits in a glass?

Host 2The Hastrup Review of Science is brought to you by ice cube trays. A low-technology solution for controlled phase transitions.

Host 1No photodetectors required. No drill hole through two kilometres of Antarctic glacier. Just water, a freezer, and patience.

Host 2They will not help you detect neutrinos. But they will keep your drink cold, which is its own form of thermal equilibrium.

Host 1Ice cube trays. For when you want a cubic-centimetre detector, not a cubic-kilometre one. Our referral link is in the show notes.

The halo profile

Host 1Both methods depend on knowing how dark matter is distributed inside the galaxy. What profile do they assume?

Host 2They use the Navarro-Frenk-White profile, called NFW, which is the standard benchmark from cosmological N-body simulations. It describes dark matter density as a function of distance from the galactic centre, with a scale radius of twenty kiloparsecs and a local density at the Sun's position of zero point four GeV per cubic centimetre.

Host 1Why does the profile matter so much?

Host 2Because the annihilation rate scales as the square of the dark matter density. The signal is proportional to a quantity called the J-factor, which is the integral of density squared along the line of sight. If the density rises steeply toward the galactic centre — as the NFW profile predicts — the J-factor is large and the limits are strong. If the inner halo is cored and the density flattens, the J-factor drops and the limits weaken.

Host 1And we do not actually know which profile is correct.

Host 2We do not. Stellar kinematics constrain the dark matter distribution in the inner Milky Way, but not tightly enough to distinguish between a cusp and a core. Simulations that include baryonic physics — the gas, stars, and feedback processes that sculpt real galaxies — give different answers depending on the implementation. The NFW profile is a reasonable central choice, but the profile uncertainty is one of the dominant systematics in any indirect dark matter search using galactic observations.

Next-generation detectors

Host 1The paper says nears freeze-out, not reaches it. What would close the gap?

Host 2Two planned instruments. IceCube-Gen2 is the next-generation neutrino telescope at the South Pole, with roughly eight times the instrumented volume of the current detector. It will collect far more shower events per year, with improved angular resolution, which directly sharpens the spatial analysis in the longitude method.

Host 1And KM3NeT.

Host 2KM3NeT is a neutrino telescope under construction in the Mediterranean Sea. Its location in the northern hemisphere gives it a direct view of the galactic centre, which is always below the horizon from that latitude. That is critical because looking through the Earth filters out the overwhelming atmospheric muon background. KM3NeT's view of the inner galaxy will be complementary to IceCube's.

Host 1And together they could cross the thermal benchmark?

Host 2The paper argues that the combination of IceCube-Gen2 and KM3NeT should push sensitivity well past the thermal relic cross section across a broad range of TeV-scale masses. If dark matter annihilates to neutrinos at the thermal rate, these detectors should see it. If they see nothing, it would substantially constrain or close off the thermal relic hypothesis for this channel.

Caveats

Host 1What should we be cautious about in this paper?

Host 2First, it is an arXiv preprint. It uses publicly available IceCube data and published effective areas, but the analysis has not gone through the IceCube collaboration's internal review or external peer review. Independent reproduction of these results would strengthen confidence substantially.

Host 1Second?

Host 2The template-dependent method, which produces the strongest constraints, inherits all the uncertainties of whatever cosmic ray model it assumes. Different templates give different limits. The paper is transparent about this — it shows results for four models — but the most optimistic claims depend on which model you consider most realistic, and that is genuinely uncertain.

Host 1Third?

Host 2The halo profile. As we discussed, the NFW profile is standard but not measured in the inner Milky Way. A cored profile would weaken the limits meaningfully. Until the inner halo shape is independently constrained, the limits carry that systematic uncertainty.

Host 1And the channel itself?

Host 2The analysis focuses exclusively on dark matter annihilating to neutrino-antineutrino pairs. This is the hardest channel to probe, which is why improving limits on it matters. But most theoretically motivated dark matter candidates do not annihilate primarily into neutrinos. The thermal relic benchmark applies to the total annihilation cross section summed over all channels. Reaching it in the neutrino channel constrains only models where neutrinos are the dominant annihilation product.

Bottom line

Host 1What should someone take away from this?

Host 2Three things. First, the detection of diffuse galactic neutrinos has opened a genuinely new avenue for dark matter searches. Before this detection, neutrino telescopes could look for dark matter in localised targets — the Sun, the galactic centre as a point source, dwarf galaxies. Having a confirmed diffuse signal across the galactic plane gives you a large, extended target where spatial and spectral analyses work simultaneously.

Host 1Second?

Host 2The thermal relic cross section — the benchmark that tells you whether dark matter was produced by freeze-out in the early universe — is now within sight for the neutrino annihilation channel. The model-independent method is within factors of a few. The template-dependent method touches it under some assumptions. This is a major advance from where the field was even three years ago.

Host 1And third?

Host 2The next decade of neutrino astronomy has a realistic path to a decisive answer. IceCube-Gen2 and KM3NeT together should either detect dark matter annihilation to neutrinos at the thermal rate or rule it out across a wide range of TeV-scale masses. That would be a landmark result in either direction — a discovery, or a firm closure of one of the last open annihilation channels at the thermal benchmark.

Sign-off

Host 1That is the deep dive. Getting Warmer: IceCube Nears Freeze Out, from Mukhopadhyay and Krnjaic, on the arXiv. Link in the show notes.

Host 2This has been the Hastrup Review of Science. See you next week.