Cosmological neutrino mass: why DESI DR2 plus CMB prefers a sum below the ≈0.06 eV oscillation floor
Difficulty 37: the culprits are identified and quantified (τ, the BAO amplitude, curvature or w0wa) and DESI DR3, Euclid and large-scale polarization data will discriminate, so it sits just below the dynamical-dark-energy and ladder problems (40), whose systematics are less isolated. Application value 8.5 is near the cosmology ceiling (dark energy 9): it informs neutrino-mass experiments and the trust placed in cosmological parameter inference. Leverage 47: the answer constrains the mass ordering and the extensions of ΛCDM, below dynamical dark energy (55) because a τ systematic would resolve it with little downstream effect.
The problem
Cosmic data say neutrinos weigh less than particle experiments require. Mistake, fluke or new physics?
- 1Oscillations: neutrinos must weigh at least 0.06 eV
- 2Cosmology: mass makes matter clump less
- 3DESI + CMB prefer less than the minimum
- 4Check optical depth, curvature, dark energy
- 5Wait for more data to settle
The oscillation floor. Neutrino oscillations fix a lower bound on the sum of neutrino masses: Σm_ν ≳ 0.058 eV for the normal mass ordering and ≳ 0.098 eV for the inverted ordering (95% CL, NuFIT 6.0).
Built from
earlier problems → this oneThe problems this one is built from: progress or tools it needs (builds on), special cases that come first (generalizes), and weaker results on the way (implies). On the right, what builds on it.
- Start with
- Step 2
- Then
- Cosmological neutrino mass: why DESI DR2 plus CMB prefers a sum below the ≈0.06 eV oscillation floorThis problem
Path to understanding
foundations → problemThe research entries this problem rests on, in the order to read them. Each links to its explainer.
- Start with
- Step 2
- Step 3
- ΛCDM, the standard model of cosmologyTheory · host
- Then
- Cosmological neutrino mass: why DESI DR2 plus CMB prefers a sum below the ≈0.06 eV oscillation floorThe problem
State of the art
DESI (Elbers et al., arXiv:2503.14744) report Σmν < 0.0642 eV (95%) for DESI DR2 BAO + CMB in ΛCDM, a Feldman–Cousins limit of 0.053 eV, and Σmν < 0.163 eV in w0waCDM; an analysis with an effective neutrino mass allowed to go negative finds a 3σ tension with the oscillation floor. NuFIT 6.0 (arXiv:2410.05380) puts that floor at 0.058 eV (normal) and 0.098 eV (inverted ordering). The leading explanations are now quantified. (i) Optical depth: dropping CMB lensing and low-ℓ polarization, the data prefer τ = 0.091 ± 0.011, which removes the tension at 0.06 eV (Jhaveri, Karwal and Hu, arXiv:2504.21813), but such a τ is in roughly 3–5σ tension with Planck low-ℓ polarization (Sailer, Farren, Ferraro and White, arXiv:2504.16932, PRL 2026); with an external τ prior of 0.11 ± 0.006 one 2026 preprint finds Σmν = 0.10 (+0.04/−0.05) eV, a tentative 2σ detection (Sullivan, de Belsunce and Ivanov, arXiv:2606.30903, unrefereed). (ii) Curvature: a small negative Ωk, favoured at about 2σ, relaxes the bound to < 0.10 eV (Chen and Zaldarriaga, arXiv:2505.00659), and with a symmetric negative-mass extension the tension falls from 2.59σ to 1.17σ, Σmν,eff = −0.011 (+0.052/−0.050) eV (Pulido-Hernández and Cervantes-Cota, arXiv:2603.13208, PRD 114, 043543, 2026). (iii) Evolving dark energy relaxes it too. No DESI DR3 or Euclid neutrino result had appeared by October 2026, so the preference is model-dependent and unresolved.
Realistic applications
A measured Σmν constrains the neutrino mass ordering and absolute mass scale, informing beta-decay and neutrinoless double-beta experiments and global cosmological parameter inference; resolving the tension also calibrates how much cosmological constraints can be trusted against systematics.
What counts as solved
Either (a) an independent determination of the dominant nuisance parameters (reionization optical depth, CMB lensing amplitude, BAO distance scale) from data not entering the preference, after which Σmν_eff is consistent with ≥ 0.059 eV at the 2σ level in ΛCDM, or (b) a model extension that fits CMB, BAO and SNe jointly with Σmν ≥ 0.059 eV with a better Bayesian evidence than ΛCDM, confirmed by DESI DR3/Euclid-class data, or (c) a >3σ detection of a negative effective neutrino energy density persisting across independent probes.
How it gets verified
ObservationSettled by new measurements or observations.
Settled by new independent measurements of the dominant nuisance parameters (reionization optical depth, CMB lensing amplitude, BAO distance scale) from data not entering the preference, after which the effective neutrino-mass sum should be consistent with 0.059 eV or more at 2 sigma in LCDM, or by a model extension that fits CMB, BAO and SNe jointly.
Claims under verification
+ add a claimNo claimed results on the Verification board for this problem yet.
Where the work is
Mostly new data (optical depth, lensing, BAO) and joint analyses; theory is needed for model extensions and consistent priors.
Bottlenecks
ranked by votes- #1EmpiricalReionization optical depth τ
Σmν is degenerate with τ through the lensing amplitude; the τ ≈ 0.09 that removes the tension conflicts with Planck low-ℓ polarization at 3–5σ, and only large-scale polarization (LiteBIRD, CLASS) or 21-cm reionization data can settle τ independently.
Community grades
Sign in to add your own grade.
Connected problems
Progress
- Mar 13, 2026Symmetric negative-mass analysis with curvaturesource
Tension with the oscillation floor falls to 1.17σ with curvature included (PRD 114, 043543).
- May 1, 2025Spatial curvature relaxes the boundsource
Chen and Zaldarriaga: a slightly negative Ωk, favoured at about 2σ, relaxes Σmν to < 0.10 eV.
- Apr 30, 2025Optical depth as the culpritsource
Jhaveri, Karwal and Hu show τ ≈ 0.09 removes the tension; Sailer et al. (2025-04-23) show this τ conflicts with Planck low-ℓ polarization.
- Mar 18, 2025DESI DR2 neutrino constraintssource
Σmν < 0.0642 eV (95%) in ΛCDM with CMB, and a 3σ tension with the oscillation floor for the effective-mass parameter.