Neutrino Mass

The question of neutrino mass unfolded across nine decades and remains unfinished. Wolfgang Pauli’s 1930 postulate treated the neutrino as massless, or nearly so, to reproduce the measured electron endpoint energy in beta decay. Fermi’s 1934 quantum theory of beta decay was compatible with any neutrino mass below a few electron-volts but had no sensitivity to smaller values. For most of the twentieth century the Standard Model was written down with three exactly massless neutrinos — a choice of mathematical economy, not of physical principle. The discovery of neutrino oscillations, announced by Super-Kamiokande in 1998 and definitively established by SNO in 2001–2002, made that choice untenable.

Neutrinos have mass. But how much? The oscillation measurements that demonstrated mass measure only the differences between squared masses, not the absolute values. Twenty-five years after oscillation was confirmed, the three neutrino masses remain unknown in absolute terms, and whether neutrinos are their own antiparticles — Dirac or Majorana — is also unknown. Three distinct experimental programmes — kinematic endpoint measurements, cosmological structure formation, and neutrinoless double beta decay — are converging on answers in the next decade.

Why oscillation implies mass

A neutrino produced as a pure flavor eigenstate — say ${\nu }_{\mu }$ from pion decay — is not an eigenstate of the free-propagation Hamiltonian unless the three mass eigenstates are exactly degenerate. In a general superposition $|{\nu }_{\alpha }⟩={\sum }_i{U}_{\alpha i}^{\ast }|{\nu }_i⟩$ each mass eigenstate propagates with its own phase $e^{-i{E}_{i}t}$, where for relativistic neutrinos $E_i\approx p+m_i^2/2p$. The phase differences $\Delta {\varphi }_{ij}=\frac{\Delta m_{ij}^2}{2E}L$ accumulate over the baseline $L$ and produce a time-dependent flavor composition — oscillation. If all $m_i$ were equal, the overall phase would factor out and no oscillation would occur.

Observation of oscillation therefore establishes that at least two of the three mass eigenstates have different, non-zero masses. The squared-mass differences measured to date are $\Delta m_{21}^2=(7.42\pm 0.20)\times 10^{-5}{\text{ eV}}^2$ $|\Delta m_{31}^2|=(2.515\pm 0.028)\times 10^{-3}{\text{ eV}}^2$ The sign of $\Delta m_{31}^2$ — whether $m_3$ is the heaviest (normal ordering) or the lightest (inverted ordering) — is still being resolved; see mass ordering. These are differences, not absolute values. Oscillation experiments alone cannot fix the mass scale.

Historical endpoint measurements

For seventy years before KATRIN, beta-decay endpoint experiments set the only model-independent upper limits on neutrino mass. The technique dates to the 1930s but did not approach sub-eV sensitivity until the end of the twentieth century. The relevant reaction is tritium decay: ${}^3\text{H}\to {}^3\text{He}+e^{-}+{\bar{\nu }}_e$ with an endpoint $E_0=18.573$ keV and half-life 12.3 years. Tritium is chosen because its low endpoint energy maximises the fractional distortion of the spectrum caused by a finite neutrino mass, and because its super-allowed decay has no nuclear-structure corrections.

Three experimental generations dominate the history:

• Curran and Angus, 1949 — first tritium endpoint using proportional counters; upper limit $m_{\nu }<1$ keV. Not tight, but the first to attempt the measurement at all.
• ITEP, 1980–1986 — the Valentin Lubimov group reported a positive detection, $m_{\nu }=30$ eV, based on careful analysis of tritium valine. This claim created major controversy and drove a decade of follow-up experiments.
• Mainz + Troitsk, 1990s–2000s — two independent experiments using electrostatic spectrometers with MAC-E filter architecture. Both set upper limits in the 2–3 eV range, definitively excluding the ITEP claim.
• KATRIN, 2019–present — Karlsruhe Tritium Neutrino, the current world-leading experiment. A 70-metre-long MAC-E spectrometer with a molecular tritium source of unprecedented purity and stability.

The 2022 and 2024 KATRIN releases pushed the bound to $m_{{\nu }_e}^{\text{eff}}<0.45\text{ eV}(90\%\text{C.L.})$ with a final design target of 0.2 eV after all anticipated systematic improvements. Project 8, a next-generation concept using cyclotron radiation emission spectroscopy (CRES) of individual tritium electrons trapped in a magnetic field, targets 0.04 eV sensitivity by the early 2030s.

What kinematic measurements actually measure

The quantity constrained by beta-decay spectra is not a single mass but an incoherent weighted sum: $m_{{\nu }_e}^{\text{eff}}=\sqrt{{\sum }_i|U_{ei}{|}^2{m}_i^2}$ In the limit of quasi-degenerate masses, all $m_i\approx m$ and this reduces to $m$. In the hierarchical limit, with $m_1$ lightest, it is dominated by $m_3$ through the coefficient $|U_{e3}{|}^2={\sin }^2{\theta }_{13}\approx 0.022$: $m_{{\nu }_e}^{\text{eff}}{|}_{\text{NO, }{m}_1=0}=\sqrt{|U_{e2}{|}^2\Delta m_{21}^2+|U_{e3}{|}^2|\Delta m_{31}^2|}\approx 9\text{ meV}$ The current KATRIN limit at 0.45 eV is therefore a factor of ~50 above the hierarchical minimum. Reaching that floor is well beyond any currently proposed experiment and will likely remain so for a generation.

Cosmological bounds

Massive neutrinos leave a distinctive imprint on the growth of cosmic structure. During the first few thousand years after the Big Bang, neutrinos are relativistic and contribute to the radiation density. They subsequently become non-relativistic and start to cluster, but their high thermal velocities allow them to “free-stream” out of forming over-densities on scales below approximately 100 Mpc.

The net effect: massive neutrinos suppress the small-scale matter power spectrum relative to a massless-neutrino universe, with the suppression amplitude scaling roughly as $\Delta P/P\approx -8f_{\nu }\text{for }k>k_{fs}$ where $f_{\nu }={\Omega }_{\nu }/{\Omega }_m$ is the neutrino fraction of the matter density. For $\sum m_{\nu }=0.1$ eV, the suppression is about 5%.

This modifies three observational signatures:

1. CMB temperature and polarisation anisotropies — through the angular scale of the first acoustic peak and the early integrated Sachs-Wolfe effect.
2. Galaxy clustering — the small-scale matter power spectrum measured by BOSS, eBOSS, DESI, and future surveys.
3. CMB lensing — the distortion of CMB photons by intervening dark-matter structure, which is sensitive to total matter including neutrinos.

Combined analyses of Planck 2018 + BOSS galaxy clustering yield (2024 compilation): ${\sum }_i{m}_{\nu }<0.12\text{ eV}(95\%\text{C.L., }\Lambda \text{CDM})$ The bound depends on the cosmological model: extending ΛCDM to allow a non-standard dark-energy equation of state, or modifying early-universe physics, can relax it by factors of a few.

This bound is already beginning to disfavor the inverted ordering, which requires $\sum m_{\nu }\gtrsim 0.10$ eV. The next generation of experiments — DESI (2021–2026), Euclid (launched 2023), Vera Rubin Observatory (first light 2025), CMB-S4 (planned) — will push the sensitivity to roughly $\sigma (\sum m_{\nu })\sim 0.02$ eV, sufficient to either definitively exclude IO or detect the minimum NO value.

An important caveat: cosmological bounds are not direct measurements of rest mass. They measure the gravitational effect of neutrinos in the cosmic expansion and structure history, convolved with assumptions about other cosmological parameters. A claimed cosmological detection of non-zero $\sum m_{\nu }$ would be a profound result but would require independent laboratory confirmation before the community would treat it as an absolute mass determination.

Neutrinoless double beta decay

If neutrinos are Majorana particles — identical to their antiparticles — they can mediate a lepton-number-violating transition in which two neutrons simultaneously convert to two protons without emitting any final-state neutrinos: $(A,Z)\to (A,Z+2)+2e^{-}$ No Standard Model process with left-handed-only neutrinos can produce this decay. The observation of neutrinoless double beta decay ($0\nu \beta \beta$) would be immediate, incontrovertible evidence that neutrinos are Majorana.

The decay rate in the “mass mechanism” is ${\Gamma }_{0\nu \beta \beta }=G_{0\nu }\cdot |M_{0\nu }{|}^2\cdot ⟨m_{\beta \beta }{⟩}^2$ where $G_{0\nu }$ is a phase-space factor (known to $\sim 1\%$), $M_{0\nu }$ is a nuclear matrix element (uncertain at the factor-of-2 level), and the effective Majorana mass is $⟨m_{\beta \beta }⟩=\left|{\sum }_i{U}_{ei}^2{m}_i\right|$ Unlike the kinematic mass, this involves coherent sums of mixing-matrix-weighted masses, with the PMNS Majorana phases entering quadratically and potentially allowing large cancellations.

Current experimental frontiers:

The ranges in $⟨m_{\beta \beta }⟩$ reflect the nuclear-matrix-element uncertainty. Next-generation experiments targeting 10-meV sensitivity — covering the full inverted-ordering parameter space even in the worst-case matrix-element scenario — are in construction or advanced design:

• LEGEND-1000 (ton-scale ⁷⁶Ge)
• nEXO (5-ton liquid-xenon TPC)
• CUPID (scintillating bolometers on ¹⁰⁰Mo)
• NEXT-100 and successors (high-pressure gas Xe)

A null result at the 10-meV level combined with a definitive mass-ordering determination favouring normal ordering would severely constrain Majorana scenarios, though not fully exclude them. A positive detection at the current ~100-meV sensitivity would be the single most important particle-physics measurement of the decade.

Why neutrino masses are so small — the seesaw

Neutrino masses are at least six orders of magnitude smaller than the next-lightest particle, the electron ($m_e=511$ keV vs. $m_{\nu }\lesssim 0.5$ eV). In the Standard Model with only left-handed neutrinos, no renormalisable mass term is allowed — leptons obtain mass through Yukawa couplings to the Higgs doublet, but the required SU(2) singlet is missing for neutrinos. Three extensions are commonly discussed.

Dirac masses require adding right-handed neutrino fields (${\nu }_R$), which must be Standard Model singlets (no gauge charges). Their Yukawa couplings to the Higgs would then be of order $10^{-12}$ to produce sub-eV masses — an unnaturally small number compared to the top-quark Yukawa of order 1, or even the electron’s $3\times 10^{-6}$. The unnaturalness is the principal reason Dirac neutrino mass is considered unattractive by many theorists.

Type-I seesaw (right-handed Majorana): introduce heavy right-handed neutrinos with Majorana mass $M_R$. The Dirac Yukawa coupling $m_D$ combines with the Majorana mass to give an effective light-neutrino mass $m_{\nu }\sim \frac{m_D^2}{M_R}$ If $m_D\sim v=174$ GeV (electroweak scale) and $M_R\sim 10^{14}$ GeV (grand-unification scale), then $m_{\nu }\sim 0.3$ eV — the right order of magnitude. The seesaw is one of the few places in physics where a specific predicted mass scale matches the observed value without fine-tuning.

Type-II seesaw (triplet Higgs): an SU(2) triplet Higgs acquires a small vacuum expectation value, giving neutrinos Majorana masses directly. The suppression of the triplet VEV relative to the electroweak scale encodes the smallness of $m_{\nu }$.

Type-III seesaw (fermion triplet): heavy SU(2) triplet fermions play the role of ${\nu }_R$. Phenomenologically similar to Type-I but with different signatures.

All three seesaw scenarios produce Majorana light neutrinos and would therefore predict $0\nu \beta \beta$. The distinction between them requires either observing heavy-sector particles (difficult unless they are at accessible scales) or precision measurements of leptonic CP violation.

Leptogenesis. Heavy right-handed neutrinos can produce a primordial lepton asymmetry in the early universe through CP-violating decays out of equilibrium. Electroweak sphalerons convert a fraction of this asymmetry into a baryon asymmetry, potentially explaining the observed matter-antimatter imbalance. The connection between low-energy PMNS CP violation and high-energy leptogenesis CP violation is model-dependent but is a central theme of modern beyond-Standard-Model physics.

Open questions and the frontier

The absolute mass scale, the Dirac-versus-Majorana question, the hierarchy (ordering), and the octant of ${\theta }_{23}$ are the four open questions that the next decade’s experiments will address:

1. Absolute scale: KATRIN (end of decade, 0.2 eV), Project 8 (early 2030s, 0.04 eV), cosmology (0.02 eV)
2. Ordering: JUNO (3σ by ~2028), DUNE + Hyper-K (5σ by early 2030s)
3. Majorana nature: LEGEND-1000, nEXO, CUPID at 10-meV sensitivity covering IO
4. CP phase: DUNE and Hyper-Kamiokande 5σ sensitivity within 5–7 years of first data

If all four come in — and the cosmological $\sum m_{\nu }$ is detected above 60 meV, the ordering is resolved, $0\nu \beta \beta$ is seen, and leptonic CP violation is confirmed — the next decade would close the largest open gap in the Standard Model since the Higgs discovery. If none comes in, the field faces a different challenge: explaining a universe whose neutrino sector is tantalisingly close to discovery but consistently just out of reach.

References

1. Fermi, E. (1934). “Versuch einer Theorie der β-Strahlen. I.” Z. Phys. 88, 161. (Foundational paper on beta decay.)
2. Super-Kamiokande Collaboration (1998). “Evidence for oscillation of atmospheric neutrinos.” Phys. Rev. Lett. 81, 1562. arXiv:hep-ex/9807003
3. SNO Collaboration (2002). “Direct evidence for neutrino flavor transformation from neutral-current interactions.” Phys. Rev. Lett. 89, 011301. arXiv:nucl-ex/0204008
4. KATRIN Collaboration (2024). “Direct neutrino-mass measurement based on 259 days of KATRIN data.” arXiv:2406.13516
5. Planck Collaboration (2020). “Planck 2018 results. VI. Cosmological parameters.” A&A 641, A6. arXiv:1807.06209
6. KamLAND-Zen Collaboration (2023). “Search for Majorana neutrinos exceeding the one-ton scale.” Phys. Rev. Lett. 130, 051801. arXiv:2203.02139
7. Minkowski, P. (1977). “μ → eγ at a rate of one out of 10⁹ muon decays?” Phys. Lett. B 67, 421. (Seesaw mechanism, Type I.)
8. Fukugita, M., Yanagida, T. (1986). “Baryogenesis without grand unification.” Phys. Lett. B 174, 45. (Leptogenesis.)

Frequently asked

How light are neutrinos?

The direct kinematic upper limit from KATRIN (2024) is m_ν < 0.45 eV at 90% confidence. Cosmological bounds on the sum of masses are tighter — roughly Σm_ν < 0.12 eV in standard ΛCDM — but model-dependent. The oscillation-imposed minimum is approximately 0.06 eV for the sum in normal ordering, or 0.10 eV in inverted ordering.

Are neutrinos their own antiparticles?

It is not yet known. If neutrinos are Majorana particles they would be identical to their antiparticles, enabling lepton-number-violating processes like neutrinoless double beta decay. Experiments like KamLAND-Zen, GERDA/LEGEND, CUORE, and nEXO are searching at increasing sensitivity; no signal has been established. Current bounds on the effective Majorana mass are ~30–150 meV depending on the nuclear matrix element used.

Why are neutrino masses so small compared to other fermions?

In the minimal Standard Model with only left-handed neutrinos, no mass term is allowed at all. Adding right-handed neutrinos produces Dirac masses requiring Yukawa couplings of order 10⁻¹², unnaturally small compared to the top-quark Yukawa of order one. The 'seesaw mechanism' instead generates small Majorana masses through a heavy right-handed partner at a new high scale (10⁹–10¹⁵ GeV), producing m_ν ~ v²/M naturally at sub-eV values.

Will cosmology resolve the mass scale before laboratory experiments?

Possibly, but with an important caveat. Forthcoming CMB (Simons Observatory, CMB-S4) and galaxy surveys (DESI, Euclid, LSST) will tighten Σm_ν bounds by roughly an order of magnitude. If a non-zero Σm_ν is detected at cosmological sensitivity — say 50 ± 20 meV — it would be a landmark result, but the measurement is model-dependent and rests on assumptions about the dark energy equation of state and other ΛCDM priors.