Dark Matter Calculations May Be Wrong When the Electroweak Transition Comes Late

For decades, physicists calculating how much dark matter should linger in today’s Universe have relied on a deceptively simple recipe: track how a dark matter particle annihilates as the cosmos cools, integrate the relevant equations from the moment it decouples from the primordial plasma, and compare the result with the abundance measured by the Planck satellite. That measurement, Omega h-squared equal to 0.1200 with an uncertainty of only 0.0012, is among the most precise numbers in cosmology, and any theoretical prediction hoping to survive must match it. But a new theoretical study published in The European Physical Journal C argues that the standard recipe contains a hidden blind spot, one that becomes dramatic precisely in the mass range where experiments are least able to test dark matter directly.

The blind spot concerns the electroweak phase transition, the moment in cosmic history when the Higgs field acquired its non-zero vacuum expectation value and endowed the fundamental particles with mass. According to the source study, this transition occurred at a temperature of roughly 160 GeV, when the Universe was a fraction of a second old and trillions of times hotter than the interior of the Sun. Before that instant, the Higgs field was switched off, particles that we now consider massive were effectively massless, and the roster of available interactions was different from what it is today. The Standard Model of particle physics, in other words, has two distinct thermal incarnations, and the transition between them rewrites the masses and couplings of everything in the plasma.

The conventional approach to computing the dark matter relic density ignores this two-phase structure. In the standard treatment, researchers solve the Boltzmann equation, which governs how the number density of dark matter particles evolves as annihilations deplete the population while the Hubble expansion dilutes it. Because dark matter is assumed to remain in thermal equilibrium until freeze-out, the integral only needs to run from the freeze-out temperature onward, and freeze-out typically happens at temperatures around one twenty-fifth of the dark matter mass. Since those temperatures are usually well below the electroweak scale, finite-temperature corrections are routinely neglected, and the calculation is performed entirely with the broken-phase model, the one describing today’s world with massive particles and a Higgs vacuum.

The flaw, as the team led by Sreemanti Chakraborti of Durham University, together with André Milagre, Rui Santos and João P. Silva of Lisbon, demonstrates, is that this logic fails for heavy dark matter. A typical scalar dark matter candidate freezes out at roughly a quarter of its mass in temperature units, so if the dark matter particle weighs more than about 4 TeV, its freeze-out occurs before the electroweak transition rather than after it. In that regime, the annihilation cross sections that determine the final abundance are evaluated in the wrong phase of the theory. The particle masses, the interaction channels and even the identity of stable particles can all differ between the unbroken and broken phases, and the study shows that plugging broken-phase quantities into the entire thermal history can shift the predicted relic density by up to factors of order one hundred percent.

To make the problem concrete, the authors constructed a minimal extension of the Standard Model scalar sector containing two real singlet fields, phi and chi, each stabilized by its own discrete Z-two symmetry. Before the electroweak transition, both scalars are stable and behave as two separate dark matter components, annihilating into pairs of Higgs doublets or into each other through portal couplings. After the transition, the Higgs doublet and the phi field both acquire vacuum expectation values, breaking the electroweak symmetry and one of the discrete symmetries. The phi particle then mixes with the Higgs boson and decays into ordinary matter, leaving chi as the sole surviving dark matter candidate, annihilating through s-channel exchange of the two Higgs-like scalars into fermions and gauge bosons.

Armed with this model, the researchers performed a random scan of one million parameter points, varying the dark matter mass between 1 and 100 TeV, the mass of the second Higgs-like scalar between 250 and 1000 GeV, the Higgs mixing angle up to 0.3, and the quartic couplings over their perturbative range. They computed the relic density twice for each point: once with the standard approach, using the broken-phase model throughout cosmic history, and once with an improved approach that switches to the unbroken-phase model above the transition temperature and the broken-phase model below it, implemented with the public code micrOMEGAs 6.0.5. The relative difference between the two results, delta Omega h-squared, served as a model-independent measure of how badly the standard calculation can go astray.

The results were striking. For dark matter masses below about 4 TeV, the two approaches agree, as expected, since freeze-out happens after the electroweak transition and the broken-phase description is valid throughout the relevant epoch. But as the mass climbs past 4 TeV, the deviation opens up sharply, and in the heavy regime above 10 TeV, where current direct detection experiments such as LUX-ZEPLIN lose sensitivity, the discrepancy can reach order one hundred percent. Crucially, the sign of the deviation can go either way. In some regions of parameter space the standard approach overestimates the abundance, wrongly excluding parameter points that the improved calculation shows are perfectly viable. In others it underestimates the abundance, admitting points that in reality produce far too much dark matter to match the Planck measurement.

The authors also provided an analytical handle on the discrepancy, decomposing it into two contributions: one capturing the shift in the yield at freeze-out, and one capturing the change in the thermally averaged annihilation cross sections across the transition. This decomposition, expressed through the quantities alpha and beta, allowed them to predict the sign of the deviation semi-analytically, and the predictions agreed excellently with the full numerical results. In one illustrative benchmark with a dark matter mass of 11.9 TeV, the improved calculation reproduced the observed relic density within three sigma, while the standard calculation missed the observed value by a staggering forty-two sigma, a difference that would lead a model builder to discard a perfectly good theory.

The physical origin of the effect is structural rather than a matter of higher-order precision. It is not simply that finite-temperature corrections shift masses and couplings slightly; rather, the very set of degrees of freedom and interaction channels available to dark matter changes across the electroweak transition. Before the transition, the dark sector in the case study contains two stable particles with a limited set of annihilation channels; after it, one particle decays away and new channels mediated by Higgs mixing open up. These phase-structure effects do not generically cancel against finite-temperature radiative corrections, which means that existing analyses of TeV-scale Higgs portal dark matter that neglect both may be systematically incomplete.

The implications reach beyond theory. Heavy dark matter candidates above 10 TeV remain largely beyond the reach of current direct detection, so relic density arguments carry enormous weight in deciding which models live or die, and a flawed calculation can therefore misdirect the entire field. The study also points toward future facilities: although the dark matter particles in these scenarios are far too heavy for any existing collider, the extended scalar sectors they inhabit could be probed at a future high-energy machine such as the FCC-hh, where the improved formalism could prove instrumental in reconstructing the underlying scalar potential. For now, the message to theorists is unambiguous: any freeze-out calculation involving a dark matter candidate heavier than about 4 TeV must reckon with the fact that the early Universe, at the crucial moment, was not yet the world we know.

Subject of Research: The impact of electroweak symmetry breaking on thermal freeze-out relic density calculations in Higgs portal dark matter models

Article Title: Interplay between electroweak symmetry breaking and Higgs portal dark matter

Article References: Interplay between electroweak symmetry breaking and Higgs portal dark matter. (n.d.). https://doi.org/10.1140/epjc/s10052-026-16295-7

Image Credits: AI Generated

DOI: 10.1140/epjc/s10052-026-16295-7

Keywords: dark matter, electroweak symmetry breaking, Higgs portal, relic density, freeze-out, Boltzmann equation, thermal freeze-out, scalar singlet models, Planck, LUX-ZEPLIN, particle cosmology, theoretical physics