Neutrinoless double beta decay

Some even-even nuclei cannot beta decay one nucleon at a time — the intermediate nucleus is heavier than the parent, so single beta decay is energetically shut. A rare few can only shed charge by converting two neutrons into two protons at once, a fourth-order weak process called double beta decay. The Standard Model version, $2\nu\beta\beta$ , emits two electrons and two antineutrinos and has been watched in a dozen isotopes with half-lives of $10^{19}$ – $10^{21}$ years — the slowest processes ever directly measured. This article is about its forbidden twin. In neutrinoless double beta decay, $0\nu\beta\beta$ , the same two neutrons turn into two protons and two electrons — but no neutrinos come out at all. The two antineutrinos that should be emitted are instead exchanged internally: one vertex's outgoing antineutrino is swallowed at the other vertex as an incoming neutrino. That trick is impossible unless the neutrino is its own antiparticle — a Majorana fermion — and unless it has mass, because the exchange demands a helicity flip that only a massive particle can supply. The process therefore violates total lepton number by two units, $\Delta L = 2$ , something no Standard Model interaction can do. This makes $0\nu\beta\beta$ the single most important experiment in neutrino physics. A confirmed signal would prove neutrinos are Majorana particles, expose the origin of their tiny masses through the seesaw mechanism, and supply the lepton-number violation that leptogenesis needs to explain why the universe is made of matter rather than nothing. The answer is hidden not in an accelerator but in tonnes of ultra-pure isotope watched for years underground.

Key takeaways

$0\nu\beta\beta$ is the decay $(A,Z)\to(A,Z+2)+2e^-$ with no neutrinos — a $\Delta L=2$ process forbidden in the Standard Model.
It can occur only if neutrinos are Majorana particles (their own antiparticles); a pure Dirac neutrino conserves lepton number and forbids it outright.
The rate factorises as $1/T_{1/2}^{0\nu}=G^{0\nu}\,|M^{0\nu}|^2\,|m_{\beta\beta}|^2/m_e^2$ — a phase-space factor, a nuclear matrix element, and the effective Majorana mass.
The effective Majorana mass $m_{\beta\beta}=|\sum_i U_{ei}^2\,m_i|$ carries Majorana phases, so it can partially cancel — vanishing is possible in the normal ordering even for Majorana neutrinos.
Best current limit (KamLAND-Zen, $^{136}$ Xe): $T_{1/2}^{0\nu}>2.3\times10^{26}$ yr, giving $m_{\beta\beta}<36$ – $156$ meV; next-generation experiments target the inverted-ordering band of $15$ – $50$ meV.
Key idea: neutrinoless double beta decay is the smoking-gun test of whether the neutrino is its own antiparticle — a nonzero rate demands a Majorana mass and lepton-number violation, which no other experiment can establish.

The core insight

Neutrinoless double beta decay converts two neutrons into two protons and two electrons with no neutrinos emitted; it can proceed only if the neutrino is a massive Majorana particle, making it the definitive laboratory test of lepton-number violation.

Where this lives today

The field is a global race of tonne-scale, ultra-low-background detectors, each betting on a different high- $Q$ isotope to beat natural radioactivity. KamLAND-Zen dissolves $^{136}$ Xe in liquid scintillator and holds the leading limit, $T_{1/2}^{0\nu}>2.3\times10^{26}$ yr, giving $m_{\beta\beta}<36$ – $156$ meV. GERDA and the Majorana Demonstrator, using enriched $^{76}$ Ge crystals, achieve the best resolution and near-zero background; their successor LEGEND-1000 aims for $T_{1/2}>10^{28}$ yr. CUORE runs a tonne of $^{130}$ Te as a cryogenic bolometer at $10$ mK, and its upgrade CUPID adds scintillation light for background rejection. nEXO proposes five tonnes of liquid $^{136}$ Xe, and SNO+ loads tellurium into the former SNO vessel. Collectively the next generation targets the full inverted-ordering band, $m_{\beta\beta}\gtrsim15$ meV, within the coming decade. The stakes reach well beyond neutrino masses. A signal would be the first laboratory proof of lepton-number violation, validating the seesaw mechanism as the origin of the neutrinos' anomalously small masses, and supplying the $\Delta L$ violation that leptogenesis needs to explain the cosmic matter–antimatter asymmetry. Through the black-box theorem, a signal by any mechanism guarantees the neutrino is a Majorana particle. Few experiments carry this much on a single monoenergetic line.

The physical picture

Picture two neutrons sitting inside a nucleus, each ready to beta decay but blocked because the one-step daughter is too heavy. They decay together. Each neutron emits a virtual $W^-$ and turns into a proton, and each $W^-$ must dispose of its charge by making an electron. The question is what happens to the neutrino line. In the ordinary $2\nu\beta\beta$ mode each $W$ launches its own antineutrino and two of them fly off — lepton number is balanced, and this is a perfectly Standard-Model process. In the $0\nu\beta\beta$ mode the neutrino emitted at one vertex is reabsorbed at the other: the internal line closes, no neutrinos escape, and the final state is two electrons carrying the entire decay energy. The first figure sets the two modes side by side and marks where the magic — and the mass insertion — has to happen.

Figure

The two double-beta modes. Left: 2νββ emits two electrons and two antineutrinos, conserves lepton number, allowed. Right: 0νββ closes the neutrino line — one antineutrino is reabsorbed as a neutrino, requiring a Majorana mass insertion ⊗ that flips helicity, so no neutrinos escape and ΔL = 2

The two modes are told apart not by counting the invisible neutrinos but by measuring the energy of the two electrons. In $2\nu\beta\beta$ the two antineutrinos steal a random share of the released energy, so the summed electron spectrum is a broad continuum from zero up to the decay $Q$ -value. In $0\nu\beta\beta$ the electrons carry everything — nuclear recoil is negligible — so their summed energy is a single sharp line pinned exactly at $Q$ . The whole experimental programme is a hunt for that one monoenergetic spike on the tail of the $2\nu$ continuum, shown in the second figure with the real $Q$ -value of the workhorse isotope $^{136}$ Xe.

Figure

Summed two-electron energy spectrum. The 2νββ mode gives a broad continuum because two antineutrinos carry away energy; the 0νββ mode piles all the energy into the electrons, giving a sharp line at the Q-value — for ¹³⁶Xe, Q = 2.458 MeV. The peak-to-tail ratio and energy resolution set the sensitivity

The core equations

The starting point is the object that makes everything possible: a mass term built from a single chirality of neutrino field. Unlike the Dirac mass, which pairs left- and right-handed fields, the Majorana mass pairs a field with its own charge conjugate, and that is only gauge-allowed because the neutrino is electrically neutral.

\mathcal{L}_M = -\tfrac12\,m_M\,\overline{\nu_L^{\,c}}\,\nu_L + \text{h.c.},\qquad \nu^c \equiv \hat C\overline{\nu}^{\,T} = \nu \quad(\text{Majorana condition}) \qquad (1)

where: $\nu_L$ — left-handed neutrino field; $\nu_L^{\,c}=\hat C\overline{\nu_L}^{\,T}$ — its charge conjugate; $m_M$ — Majorana mass; the reality condition $\nu^c=\nu$ says the particle is its antiparticle. Reading it: this term carries lepton number $\Delta L=2$ , so a nonzero $m_M$ is exactly what breaks lepton-number conservation. The measurable consequence is a decay rate. Because the leptonic physics (neutrino exchange) and the hadronic physics (the nuclear transition) live at very different scales, the inverse half-life factorises cleanly into three independent pieces.

\frac{1}{T_{1/2}^{0\nu}} = G^{0\nu}(Q,Z)\,\bigl|M^{0\nu}\bigr|^2\,\frac{|m_{\beta\beta}|^2}{m_e^2} \qquad (2)

where: $T_{1/2}^{0\nu}$ — the $0\nu\beta\beta$ half-life; $G^{0\nu}$ — phase-space factor, exactly calculable, scaling as $\sim Q^5$ ; $M^{0\nu}$ — dimensionless nuclear matrix element from nuclear-structure theory; $m_{\beta\beta}$ — effective Majorana mass; $m_e$ — electron mass. Reading it: the particle-physics unknown $m_{\beta\beta}$ is bracketed by a clean number $G^{0\nu}$ and a hard-to-compute number $M^{0\nu}$ , so extracting a mass from a half-life is only as good as the nuclear theory. The particle-physics content sits entirely in the effective Majorana mass — a coherent sum over the three neutrino mass eigenstates weighted by the squares (not moduli-squared) of the first-row PMNS elements.

m_{\beta\beta} = \left|\sum_{i=1}^{3} U_{ei}^2\,m_i\right| = \left|\,c_{12}^2c_{13}^2\,m_1 + s_{12}^2c_{13}^2\,m_2\,e^{i\alpha_2} + s_{13}^2\,m_3\,e^{i\alpha_3}\right| \qquad (3)

where: $U_{ei}$ — electron-row PMNS elements; $m_i$ — neutrino masses; $c_{ij},s_{ij}=\cos\theta_{ij},\sin\theta_{ij}$ ; $\alpha_2,\alpha_3$ — Majorana CP phases, physical only for Majorana neutrinos. Reading it: because the terms are complex, they can add or cancel — this is why $m_{\beta\beta}$ is not simply the neutrino mass scale. Finally, the reason a Majorana mass rather than mere Majorana character is required: the exchanged neutrino must flip helicity between the two vertices, and the amplitude for that flip is proportional to $m_\nu/E$ .

\mathcal{A}_{0\nu} \;\propto\; \sum_i U_{ei}^2\,\frac{m_i}{q^2 - m_i^2} \;\xrightarrow{\ m_i^2\ll q^2\ }\; \frac{1}{q^2}\sum_i U_{ei}^2\,m_i = \frac{m_{\beta\beta}}{q^2} \qquad (4)

where: $q$ — momentum transferred through the virtual neutrino, of order the nuclear Fermi momentum $\sim100$ MeV $\gg m_i$ ; the numerator $m_i$ is the helicity-flip factor from the propagator. Reading it: the rate $\propto|\mathcal{A}|^2\propto m_{\beta\beta}^2$ , so a heavier neutrino means a faster decay — and a massless neutrino means no decay at all.

Derivation AWhy the decay needs a massive Majorana neutrino

The goal is to show, from the chiral structure of the weak charged current alone, that closing the neutrino line in double beta decay is impossible unless two conditions both hold: the neutrino equals its own antiparticle, and it carries a nonzero mass. We follow the neutrino from the vertex where it is created to the vertex where it must be absorbed, track its lepton number and its helicity at each step, and find that each is wrong by exactly the amount a Majorana mass insertion can repair. This single argument explains why $0\nu\beta\beta$ is simultaneously a test of lepton-number violation and a probe of the absolute neutrino mass scale.

Step 1. At the first neutron's vertex the charged current $\bar u\gamma^\mu(1-\gamma^5)d$ paired with $\bar e\gamma_\mu(1-\gamma^5)\nu$ produces, alongside the electron, a right-handed antineutrino — that is what the $V-A$ interaction emits. At the second vertex, to make the second electron, the same current must absorb a left-handed neutrino. Why: the weak interaction is chiral, so the two vertices demand opposite lepton number (antineutrino out versus neutrino in) and opposite helicity.

Step 2. Reconcile the lepton number. The line can only close if the object emitted as an antineutrino is the very object absorbed as a neutrino — i.e. if antineutrino and neutrino are the same particle. That is the Majorana condition $\nu^c=\nu$ ; for a Dirac neutrino the two are distinct states and the diagram simply does not exist. Why: a Dirac neutrino carries a conserved lepton number that forbids the $\Delta L=2$ topology outright, so Majorana character is a strict prerequisite.

Step 3. Reconcile the helicity. Even for a Majorana particle the state emitted is dominantly right-handed while the state needed is left-handed; the two overlap only through the "wrong-helicity" component of a massive spinor, whose amplitude is $\sim m_\nu/E$ . A strictly massless neutrino has definite helicity and zero overlap. Why: helicity is a good quantum number only in the massless limit, so the mismatch is cured precisely by the mass, giving amplitude $\propto m_\nu$ .

Step 4. Assemble. The propagator numerator therefore supplies one power of the neutrino mass, summed coherently over eigenstates with the PMNS weights of equation (4), yielding $\mathcal{A}\propto m_{\beta\beta}/q^2$ and a rate $\propto m_{\beta\beta}^2$ . Why: this is the "mass mechanism" — the standard interpretation of a $0\nu\beta\beta$ signal as a direct measurement of a lepton-number-violating neutrino mass. *Why both limits matter: the rate vanishes as $m_\nu\to0$ and for a pure Dirac neutrino — both limits reproduce the Standard-Model statement that $0\nu\beta\beta$ is forbidden.*

Go deeper · Derivation B · 2 minFactorising the half-life into phase space, nuclear physics, and mass

Derivation BFactorising the half-life into phase space, nuclear physics, and mass

The aim is to justify why equation (2) splits so cleanly into three factors, and to see what each one hides. Double beta decay is a fourth-order weak process inside a nucleus, and a brute-force calculation would tangle nuclear wavefunctions with lepton kinematics hopelessly. The saving grace is a separation of scales: the exchanged neutrino carries momentum of order the nuclear Fermi scale, far above the neutrino masses but far below the electroweak scale, letting the leptonic and hadronic pieces be evaluated almost independently and multiplied.

Step 1. Start from Fermi's golden rule for the rate, $\Gamma = 2\pi\,\overline{|\mathcal{M}|^2}\,\rho_f$ , with the matrix element built from two charged-current vertices joined by the light-neutrino propagator of equation (4). Factor the amplitude into a leptonic current times a nuclear current. Why: the two currents involve disjoint degrees of freedom (electrons and neutrino versus nucleons), so the amplitude is a product before any integration.

Step 2. Integrate the two outgoing electrons over their phase space. Because their summed energy is fixed at $Q$ , this integral is a pure kinematic number depending only on $Q$ and the daughter charge $Z$ (through the Coulomb distortion of the electron waves): the phase-space factor $G^{0\nu}(Q,Z)\propto Q^5$ . Why: fixing the electron energy sum removes all nuclear dependence from the leptonic integral, isolating a calculable factor that rewards high- $Q$ isotopes.

Step 3. Collapse the nuclear part. The neutrino propagator, integrated over its momentum, becomes a neutrino potential connecting the two decaying nucleons; sandwiched between initial and final nuclear states it defines the dimensionless matrix element $M^{0\nu}=\langle Z+2|\,\mathcal{O}_{0\nu}\,|Z\rangle$ . Why: the long-range neutrino potential makes the two nucleons decay coherently, and all the many-body nuclear-structure difficulty is packed into this one number, uncertain by a factor of two to three across methods.

Step 4. Pull the leptonic mass factor out front. The propagator numerator contributes the common factor $m_{\beta\beta}/m_e$ (normalised by $m_e$ to keep $M^{0\nu}$ dimensionless), leaving equation (2). Why: the particle-physics unknown thus multiplies a product of a known ( $G^{0\nu}$ ) and a modelled ( $M^{0\nu}$ ) factor, so the map from measured $T_{1/2}$ to inferred $m_{\beta\beta}$ inherits the full nuclear-matrix-element uncertainty. *Dimensional check: $[G^{0\nu}]=$ yr $^{-1}$ , $M^{0\nu}$ and $m_{\beta\beta}/m_e$ are dimensionless, so $1/T_{1/2}$ correctly carries units of yr $^{-1}$ .*

Go deeper · Derivation C · 1 minThe effective mass, Majorana phases, and cancellation

Derivation CThe effective mass, Majorana phases, and cancellation

The purpose here is to understand why the quantity probed by $0\nu\beta\beta$ is the peculiar combination $m_{\beta\beta}=|\sum U_{ei}^2 m_i|$ rather than any single neutrino mass, and why it can be far smaller than the masses themselves — even zero. The key is that the neutrino propagating between the two vertices is not a mass eigenstate but the electron-flavour state, a coherent superposition of $\nu_1,\nu_2,\nu_3$ whose relative phases matter.

Step 1. Expand the electron-flavour neutrino in mass eigenstates, $\nu_e=\sum_i U_{ei}\nu_i$ . Both the emission (as $\bar\nu_e$ ) and the absorption (as $\nu_e$ ) bring one factor of $U_{ei}$ , so each eigenstate enters the amplitude with $U_{ei}^2$ — the square, not the modulus squared. Why: because the same flavour state appears at both vertices without complex conjugation, the physical Majorana phases in $U_{ei}$ survive rather than cancelling as they would in oscillations.

Step 2. Attach the mass factor $m_i$ from the helicity flip (Derivation 1) and sum, giving $m_{\beta\beta}=|\sum_i U_{ei}^2 m_i|$ . Insert the standard parametrisation with two extra phases $\alpha_2,\alpha_3$ to reach equation (3). Why: these Majorana phases are unobservable in oscillation experiments but real here, so $0\nu\beta\beta$ accesses information no oscillation measurement can.

Step 3. Read off the consequence. In the normal ordering with a small lightest mass, the three terms are comparable in size but can carry opposing phases, so they can partially or completely cancel — $m_{\beta\beta}$ has a "funnel" where it dips to zero. In the inverted ordering the two dominant terms share the near-equal atmospheric mass and cannot fully cancel, leaving a floor near $15$ meV. Why: this asymmetry between orderings is why the inverted ordering is the natural first target — it guarantees a nonzero, reachable signal. *As a consistency check: with $m_1=0$ the normal ordering gives $m_{\beta\beta}\in[1.4,3.7]$ meV; with $m_3=0$ the inverted ordering gives $[19,49]$ meV — both ranges are worked out in detail in the examples below.*

Limiting cases

Case	Result	Meaning
Dirac neutrino ( $m_M=0$ )	$m_{\beta\beta}=0$ , rate $=0$	lepton number conserved; $0\nu\beta\beta$ strictly forbidden
Massless neutrino	$\mathcal{A}\propto m_\nu\to0$	helicity flip impossible; no decay even if Majorana
Normal ordering, $m_{\text{lightest}}\to0$	$m_{\beta\beta}\approx1$ – $4$ meV	terms comparable; Majorana phases can cancel to $\approx0$ (the funnel)
Inverted ordering, $m_{\text{lightest}}\to0$	$m_{\beta\beta}\approx15$ – $50$ meV	atmospheric terms dominate, no full cancellation — next-gen target
Quasi-degenerate ( $m_i\gg\sqrt{\Delta m^2}$ )	$m_{\beta\beta}\to m_0\,\|{\textstyle\sum}U_{ei}^2\|$	approaches the common mass $m_0$ ; tens–hundreds of meV, already constrained
Heavy Majorana exchange (seesaw $N$ )	$\mathcal{A}\propto1/M_N$	same $\Delta L=2$ operator via heavy neutrinos; a distinct mechanism

Common trapMeasuring a neutrino mass proves it is Majorana

It does not. Oscillation experiments prove neutrinos have nonzero mass differences; the KATRIN endpoint experiment and cosmology bound the mass scale. None of these says anything about whether the neutrino is Dirac or Majorana — they are all blind to lepton number. A Dirac neutrino can be as massive as you like and still forbid $0\nu\beta\beta$ . Only a genuine $\Delta L=2$ observation settles the question, and $0\nu\beta\beta$ is the one such process within experimental reach. The converse trap is just as important: a null $0\nu\beta\beta$ result never disproves the Majorana nature, because $m_{\beta\beta}$ can vanish by phase cancellation in the normal ordering even when neutrinos are Majorana. Absence of the signal constrains $m_{\beta\beta}$ , not the nature of the neutrino.

Worked examples

1 · The inverted-ordering band (fully worked)

Assume the inverted mass ordering with a vanishing lightest state, $m_3=0$ . Using $\sin^2\theta_{12}=0.304$ , $\sin^2\theta_{13}=0.0222$ , $\Delta m^2_{21}=7.4\times10^{-5}\ \mathrm{eV}^2$ and $|\Delta m^2_{31}|=2.5\times10^{-3}\ \mathrm{eV}^2$ , find the allowed range of $m_{\beta\beta}$ as the Majorana phases vary.

Step 1. Fix the masses. In the inverted ordering $m_1$ and $m_2$ sit near the atmospheric scale and $m_3$ is smallest. With $m_3=0$ : $m_1\approx\sqrt{|\Delta m^2_{31}|}=\sqrt{2.5\times10^{-3}}=\mathbf{0.050\ eV}$ , and $m_2=\sqrt{m_1^2+\Delta m^2_{21}}\approx0.0501$ eV — the two are equal to better than a percent, which is what forbids full cancellation.

Step 2. Collect the mixing weights: $c_{13}^2=1-0.0222=0.978$ , $c_{12}^2=1-0.304=0.696$ , $s_{12}^2=0.304$ . The $m_3$ term drops out ( $m_3=0$ ), so $m_{\beta\beta}=c_{13}^2\,m_1\,\bigl|\,c_{12}^2 + s_{12}^2\,e^{i\alpha}\,\bigr|$ where $\alpha$ is the surviving relative Majorana phase.

Step 3. Extremise over $\alpha$ . The bracket is largest when the terms align ( $\alpha=0$ ): $|c_{12}^2+s_{12}^2|=1$ . It is smallest when they oppose ( $\alpha=\pi$ ): $|c_{12}^2-s_{12}^2|=|0.696-0.304|=0.392$ .

Step 4. Multiply through with $c_{13}^2 m_1 = 0.978\times0.050 = 0.0489$ eV. Maximum: $m_{\beta\beta}=0.0489\times1.0=\mathbf{48.9\ meV}$ . Minimum: $m_{\beta\beta}=0.0489\times0.392=\mathbf{19.2\ meV}$ . The inverted ordering therefore predicts $m_{\beta\beta}\approx19$ – $49$ meV — a floor that cannot dip to zero, which is exactly why next-generation experiments are designed to cover this band.

2 · The normal-ordering funnel

Repeat for the normal ordering with $m_1=0$ , and show that cancellation to nearly zero is possible.

Step 1. Masses: $m_1=0$ , $m_2=\sqrt{\Delta m^2_{21}}=\sqrt{7.4\times10^{-5}}=8.6\times10^{-3}$ eV, $m_3=\sqrt{\Delta m^2_{31}}=0.050$ eV.

Step 2. The three terms of equation (3): the $m_1$ term vanishes; $s_{12}^2c_{13}^2 m_2 = 0.304\times0.978\times0.0086=\mathbf{2.56\ meV}$ ; $s_{13}^2 m_3 = 0.0222\times0.050=\mathbf{1.11\ meV}$ .

Step 3. These two surviving terms are comparable in size. Aligned, $m_{\beta\beta}=2.56+1.11=\mathbf{3.67\ meV}$ . Opposed, $m_{\beta\beta}=|2.56-1.11|=\mathbf{1.45\ meV}$ . Allowing a slightly nonzero $m_1$ lets the $m_1$ term grow and knife the remainder down to exactly zero — the normal-ordering funnel.

Step 4. The lesson: the normal ordering spans $m_{\beta\beta}\sim0$ – $4$ meV near $m_1=0$ , an order of magnitude below the inverted band and below the reach of any planned experiment. A null result at $15$ meV would exclude the inverted ordering (for Majorana neutrinos) but say nothing about the normal ordering.

3 · From a half-life limit to a mass limit

KamLAND-Zen reports $T_{1/2}^{0\nu}(^{136}\mathrm{Xe})>2.3\times10^{26}$ yr. Using a phase-space factor $G^{0\nu}=1.5\times10^{-14}\ \mathrm{yr}^{-1}$ and a representative nuclear matrix element $|M^{0\nu}|=2.7$ , estimate the bound on $m_{\beta\beta}$ .

Step 1. Invert equation (2): $m_{\beta\beta}=m_e\big/\sqrt{T_{1/2}^{0\nu}\,G^{0\nu}\,|M^{0\nu}|^2}$ , with $m_e=5.11\times10^5$ eV.

Step 2. Evaluate the denominator: $T_{1/2}G^{0\nu}|M^{0\nu}|^2 = (2.3\times10^{26})(1.5\times10^{-14})(2.7^2)=2.5\times10^{13}$ , whose square root is $5.0\times10^{6}$ .

Step 3. Divide: $m_{\beta\beta}<5.11\times10^5/5.0\times10^6=\mathbf{0.10\ eV}\approx100\ meV$ .

Step 4. Note the nuclear caveat: matrix elements from $1.6$ to $4.8$ move the bound from $\sim175$ meV to $\sim58$ meV, which is why the published limit is a range, $m_{\beta\beta}<36$ – $156$ meV — the spread is nuclear theory, not statistics. The measurement already probes the top of the inverted-ordering band; reaching $15$ meV needs another factor of $\sim50$ in half-life.

Physics of the portrait

$0\nu\beta\beta$ is the experimental shadow of an idea one man wrote down in his last published paper: that a neutral fermion can be built to be its own antiparticle.

Origins · 1 minEttore Majorana and how this was discovered

Ettore Majorana

1906–1938 · the real fermion field and neutrino self-conjugacy

Ettore Majorana was, by Enrico Fermi's own account, a physicist of the rank of Galileo and Newton — and one who published almost nothing and then vanished. Born in Catania in 1906, he joined Fermi's Rome group ("the Via Panisperna boys") and worked with an intensity that produced deep results he often declined to write up, judging them too obvious or not yet perfect. His decisive contribution came in 1937, in what would be his final paper, "Teoria simmetrica dell'elettrone e del positrone": he showed that a neutral spin-½ field can satisfy a reality condition, $\psi=\psi^c$ , so that the particle and its antiparticle are one and the same — the Majorana fermion. For charged particles this is impossible, but for a neutral one, such as the neutrino, it is allowed, and it changes the neutrino's mass structure and its lepton-number bookkeeping completely. In March 1938 Majorana withdrew his savings, boarded a ship between Palermo and Naples, and disappeared without trace. Eight decades later, tonnes of xenon and germanium sit in deep underground laboratories testing the one prediction his idea makes for the neutrino: that lepton number is not sacred, and that two electrons can be born where none should be.

History

1935

— Goeppert-Mayer computes the rate of two-neutrino double beta decay as a second-order weak process

1937

— Majorana formulates the symmetric theory in which a neutral fermion is its own antiparticle

1939

— Furry proposes neutrinoless double beta decay via Majorana neutrino exchange

1982

— Schechter and Valle prove the black-box theorem linking any 0νββ signal to a Majorana mass

1987

— Elliott, Hahn and Moe make the first direct laboratory observation of 2νββ in selenium-82

Going further

Texts. Thomson, Modern Particle Physics, §17.8 (neutrino masses) and §17.8.1 (the seesaw mechanism) · Griffiths, Introduction to Elementary Particles, 2nd ed., Ch. 11 (neutrino oscillations and mass) · Zuber, Neutrino Physics, 3rd ed., Ch. 7 (double beta decay) · Giunti & Kim, Fundamentals of Neutrino Physics and Astrophysics, Ch. 14 · Schwartz, Quantum Field Theory and the Standard Model, §11 (Majorana fermions). Oscillation parameters use the current global-fit central values ( $\sin^2\theta_{12}=0.304$ , $\sin^2\theta_{13}=0.0222$ , $\Delta m^2_{21}=7.4\times10^{-5}$ eV², $|\Delta m^2_{31}|=2.5\times10^{-3}$ eV²); the phase-space factor and representative nuclear matrix element follow Kotila–Iachello and standard shell-model/QRPA compilations.

Prerequisites. Majorana vs Dirac neutrinos · Neutrino mass mechanisms · Chirality and Weyl fields · The weak charged current. Leads to. Neutrino mass beyond the Standard Model · Matter–antimatter asymmetry and leptogenesis · Grand unification.

Ask the tutor: Ask me to walk through the helicity-flip argument line by line, or to plot how $m_{\beta\beta}$ traces the lobster-shaped band as the lightest neutrino mass and Majorana phases vary.

Why does observing 0νββ prove the neutrino is Majorana?

Because of the Schechter–Valle "black-box" theorem (1982). Whatever new physics actually drives the decay — light-neutrino exchange, a heavy right-handed neutrino, supersymmetric particles, or something else — the very existence of a $\Delta L=2$ amplitude connecting two neutrons to two protons and two electrons can always be closed into a diagram that generates a Majorana mass term for the neutrino. You cannot have the decay without also having $\nu=\bar\nu$ at some level. So the interpretation of the rate (which mechanism, what $m_{\beta\beta}$ ) may be debated, but the qualitative conclusion — neutrinos are Majorana — is guaranteed by any nonzero signal.

Doesn't KATRIN or cosmology already tell us the neutrino mass?

They tell you the mass scale, not the nature. KATRIN measures the endpoint of tritium beta decay and bounds the effective electron-neutrino mass $m_\beta=\sqrt{\sum|U_{ei}|^2 m_i^2}$ ; cosmology bounds the sum $\sum m_i$ through structure formation. Both are completely insensitive to lepton number and to the Majorana phases, so neither can distinguish a Dirac from a Majorana neutrino. The three quantities $m_\beta$ , $\sum m_i$ and $m_{\beta\beta}$ are different weighted combinations of the same masses — and only $m_{\beta\beta}$ , accessed by $0\nu\beta\beta$ , is nonzero exclusively for Majorana neutrinos.

Why is the nuclear matrix element such a headache?

The matrix element $M^{0\nu}$ is the overlap of the parent and daughter nuclear ground states connected by a two-nucleon operator with a long-range neutrino potential — a genuine quantum many-body problem for nuclei with dozens of nucleons. Different methods (shell model, QRPA, energy-density functionals, ab initio) disagree by factors of two to three, and since the rate goes as $|M^{0\nu}|^2$ , that becomes a factor of four to nine in the extracted $m_{\beta\beta}^2$ . This is why limits are quoted as ranges, and why measuring the same $m_{\beta\beta}$ in several isotopes is prized as a cross-check on the nuclear theory.

Could a null result ever rule out Majorana neutrinos?

No — and this is a subtle but crucial point. In the normal mass ordering the three contributions to $m_{\beta\beta}$ can cancel through their Majorana phases, driving the effective mass to zero even though the neutrino is genuinely Majorana. So an experiment that sees nothing has only placed an upper limit on $m_{\beta\beta}$ ; it cannot conclude the neutrino is Dirac. A positive signal proves Majorana; a negative one constrains a mass and, if deep enough, can exclude the inverted ordering, but it can never falsify the Majorana hypothesis outright.

Why these particular isotopes — xenon, germanium, tellurium?

Three demands drive the choice. First, a high $Q$ -value: the rate scales roughly as $Q^5$ , and a $Q$ above $2$ MeV lifts the signal line clear of most natural gamma backgrounds. Second, single beta decay of the parent must be energetically forbidden, so double beta decay is the only channel. Third, the isotope must be enrichable and deployable in radiopure tonne quantities. $^{136}$ Xe ( $Q=2.458$ MeV), $^{76}$ Ge ( $2.039$ MeV), $^{130}$ Te ( $2.527$ MeV) and $^{100}$ Mo ( $3.034$ MeV) all thread these needles, each trading resolution against background against cost.

How does 0νββ connect to why the universe has matter at all?

Through leptogenesis. The heavy right-handed Majorana neutrinos of the seesaw can decay in the early universe in a way that violates lepton number and CP, producing an excess of leptons over antileptons; Standard-Model "sphaleron" processes then reprocess part of that lepton asymmetry into the baryon asymmetry we observe. The essential ingredient is Majorana mass — lepton-number violation — which is precisely what $0\nu\beta\beta$ tests directly. A signal would not measure the cosmic asymmetry, but it would confirm that nature possesses the one mechanism leptogenesis cannot do without.