← quantum · contents

§ Phase xi · Mass & Scale · Chapter 03

Why neutrinos are light

On a December evening in 1930, a Viennese physicist sat down to write a letter beginning 'Dear Radioactive Ladies and Gentlemen,' and invented a particle he was sure would never be seen. Ninety-six years later it remains the lightest matter particle anyone has ever weighed, and the only one whose mass nobody can yet explain.

The letter was dated 4 December 1930. Wolfgang Pauli was thirty years old, living in Zürich, and had just sent his apologies to a meeting of radioactivity specialists in Tübingen because, as he explained, he could not leave a Christmas ball he had promised to attend. In place of attendance he sent a typed manifesto, addressed in mock-formal style to “Dear Radioactive Ladies and Gentlemen,” and asked the chair to read it aloud. The letter contained, with characteristic Pauli bluntness, the worst confession a theoretical physicist can make. “I have hit upon a desperate remedy,” he wrote, “to save the law of conservation of energy.”

The problem he was trying to rescue was beta decay. When a radioactive nucleus spits out an electron, the electron flies off with a perfectly definite kinetic energy if you believe the textbook. The parent and daughter nuclei are both at rest, both have well-defined masses, and what is left over has to come out as the electron’s energy of motion. Subtract the daughter from the parent, multiply by c², and the result is what the electron should carry. That is what conservation of energy demands.

But the electrons of beta decay refused to play along. James Chadwick had shown in 1914, with painstaking measurements at the Physikalisch-Technische Reichsanstalt in Berlin, that they emerged with a continuous spread of energies, anywhere from almost zero up to a sharp maximum. The maximum sat exactly where conservation of energy predicted; but every other electron came out short. Energy was, apparently, missing. By 1929 the situation was so bad that Niels Bohr, who had built his career on quantum strangeness, was prepared to abandon energy conservation altogether in the subatomic realm. He proposed that energy was conserved only statistically, on average, over many decays. Pauli refused to accept this. He thought it sloppy. The remedy in his Tübingen letter was less sloppy and almost as outrageous: a new particle, hiding inside the nucleus, neutral in charge, lighter than the electron, escaping every detector then known, and carrying off the energy the electron lacked. He called it a “neutron.” Two years later James Chadwick discovered something much heavier and stole the name. Enrico Fermi, half in jest, half because his Italian colleagues kept asking what to call the missing thing, suggested the diminutive: neutrino, the little neutral one.

For twenty-six years the neutrino sat in physics like a placeholder. Fermi built it into his elegant 1934 theory of beta decay, the first quantum field theory of the weak force, where the neutrino played the role of a third dancer carrying off momentum and energy in every decay. Hans Bethe and Rudolf Peierls calculated in 1934 what it would take to detect one and came to a discouraging answer: a neutrino could cross light-years of lead before bumping into anything. They wrote in the journal Nature that “there is no practically possible way of observing the neutrino.” Bethe, in his sober Cambridge style, added, “this is no real difficulty.” It was real enough; it was an embarrassment. Pauli was reported to have offered a case of champagne to anyone who detected his ghost.

The case was claimed in 1956. Frederick Reines and Clyde Cowan, two young Americans working at the Los Alamos National Laboratory, had a brilliantly bad idea: detect the neutrino’s antiparticle, the antineutrino, using a nuclear reactor as a deliberate source. A fission reactor at Savannah River in South Carolina pours out something on the order of 10^20 antineutrinos every second, and although the cross-section is absurd (the typical antineutrino has only a one-in-a-quadrillion chance of interacting in a metre of water), 10^20 is a large number to multiply against. Reines and Cowan built two thousand-litre tanks of cadmium-doped water, surrounded them with photomultiplier tubes, parked the apparatus at the reactor, and waited for the coincident flash-bang of a positron annihilation followed microseconds later by a neutron capture. They saw it. They wired Pauli the news. Pauli sent the champagne.

The Nobel Prize for that work came in 1995, almost forty years late, by which time Cowan had been dead for twenty-one years and could not share it. Pauli himself had died in 1958, two years after the detection, of pancreatic cancer in a Zürich hospital. The room number in which he died, the friends he received afterward gleefully noted, was 137, the inverse of the fine-structure constant he had spent his life trying to derive.

The detection at Savannah River was only the prologue. The deep mystery, the one that ended up rewriting the Standard Model, came from a smaller, more stubborn problem: when you watched the Sun, the neutrinos arrived in the wrong numbers and, eventually, in the wrong flavours.

In the 1960s, the now-famous Homestake experiment made the first measurement of the flux of electron neutrinos arriving from the core of the Sun and found a value that was between one third and one half the number predicted by the Standard Solar Model. This discrepancy, which became known as the solar neutrino problem, remained unresolved for some thirty years, while possible problems with both the experiment and the solar model were investigated, but none could be found. Eventually, it was…

To see why this mattered, you have to know what physicists thought neutrinos were in 1990. There were three of them, one for each charged lepton: the electron neutrino ν_e (Pauli’s original, the one that comes out of beta decay), the muon neutrino ν_μ (discovered in 1962 at Brookhaven by Lederman, Schwartz, and Steinberger by aiming an accelerator beam at a steel pile and watching what came out), and the tau neutrino ν_τ (inferred from missing energy in tau-lepton decays in the late 1970s and directly detected by the DONUT collaboration at Fermilab only in 2000). All three were massless in the Standard Model, by assumption, because no experimental evidence required otherwise and the simplest version of the theory worked beautifully without giving them mass. The neutrinos sat in the model as the only matter particles excused from the Higgs mechanism, three little zeros at the bottom of the mass table.

The Sun makes electron neutrinos. The fusion of four protons into one helium nucleus, the reaction that powers every solar-type star, releases two positrons and two electron neutrinos for every helium nucleus produced. The total flux at Earth is enormous: about 6 × 10^10 solar neutrinos per square centimetre per second pass through your fingernail right now, by day or by night (the Earth is no obstacle). Ray Davis at Homestake had been looking specifically for the electron-neutrino component, since his chlorine-to-argon reaction was sensitive only to ν_e. He kept finding about one-third of the predicted rate. John Bahcall, the Princeton astrophysicist who had calculated the prediction, kept refining his solar models. The deficit refused to budge.

Two experiments cracked the problem at the turn of the millennium. The first was Super-Kamiokande, a 50,000-tonne tank of ultrapure water buried a kilometre under Mount Ikenoyama in Japan, surrounded by 11,146 photomultiplier tubes. It detected neutrinos by watching for the faint blue Cherenkov cones produced when an incoming neutrino kicked an electron into faster-than-light-in-water motion. In 1998 the Super-K collaboration announced that atmospheric neutrinos, the ones produced when cosmic rays smash into the upper atmosphere and make pions that decay into muons and then electrons, showed a disturbing asymmetry. Neutrinos coming down from overhead arrived in the expected ratios. Neutrinos coming up through the entire Earth (produced on the far side of the planet and travelling through the core to reach Tokyo from below) showed a deficit of muon flavours. The Earth was not eating them. Something else was. The conclusion, delivered by the collaboration spokesperson Takaaki Kajita at the Neutrino 98 conference in Takayama and met with a long silence followed by a standing ovation, was that muon neutrinos were oscillating into something else, almost certainly tau neutrinos, on the way through the planet. Oscillation required mass. The Standard Model had been quietly wrong for thirty years.

The second experiment was SNO, the Sudbury Neutrino Observatory, a thousand tonnes of heavy water (D₂O, on loan from the Canadian nuclear reserve) suspended in a Plexiglas sphere two kilometres underground in an Ontario nickel mine. The deuterium in heavy water gave SNO a trick that ordinary water could not pull off. SNO could count solar neutrinos in two ways. One reaction was sensitive only to electron neutrinos, the same trick Davis had used: ν_e + d → p + p + e⁻. The other reaction, neutral-current dissociation of the deuteron, was sensitive to all three flavours equally: ν + d → n + p + ν. If the Sun produced only ν_e and they arrived intact, the two counts would agree. If the ν_e was being converted into ν_μ or ν_τ in transit, the flavour-blind count would be larger than the flavour-sensitive count, by exactly the missing factor of three Davis had been measuring for thirty years.

In June 2001, with Arthur McDonald announcing the result in Sudbury, SNO published the answer. The total neutrino flux from the Sun, summed across all three flavours, agreed with John Bahcall’s solar-model prediction to about five percent. The electron-only flux was a third of that total. Two-thirds of the solar electron neutrinos had been converted into muon and tau neutrinos somewhere on the eight-minute trip from the solar core to the Plexiglas sphere. The Sun was working as predicted; the neutrinos had been lying. Or rather, the Standard Model had. Davis and Kajita and McDonald split the 2002 and 2015 Nobels for the work. Bahcall died in 2005, weeks after the SNO confirmation was finally beyond statistical doubt.

Oscillation gave us the differences between neutrino masses, encoded in two splittings: a large one Δm²_atm ≈ 2.5 × 10⁻³ eV², set by the atmospheric data, and a smaller one Δm²_sol ≈ 7.5 × 10⁻⁵ eV², set by the solar data. The squared mass differences, translated into the kind of mass quantity Pauli would have recognised, sit in the milli-eV regime: a few hundredths of an eV at most, separating the three mass states. What oscillation does not give us is the absolute mass scale, only the splittings. The lightest neutrino could weigh almost nothing; the heaviest, the upper splitting tells us, weighs at least 0.05 eV. Direct laboratory measurements of the electron-neutrino mass, the KATRIN experiment in Karlsruhe being the current champion, set a model-independent upper bound of about 0.8 eV. Cosmological observations (the way neutrino mass tugs at the formation of large-scale structure in the universe) push that bound down to about 0.1 eV for the sum of all three. Compare to the electron, the next-lightest matter particle, weighing 511,000 eV. The neutrino is at least a million times lighter than anything else in the standard table.

The mathematics is the punchline. The reason oscillations require mass is so clean it could fit on a postcard. A massless particle travels at the speed of light and ages not at all (its own proper time is zero). A particle that does not age cannot oscillate, because oscillation requires a phase that evolves. As soon as you find evidence of a neutrino changing flavour mid-flight, you have, in one stroke, proven it has mass, and you have also handed yourself a clock: the rate at which the oscillation cycles tells you how big the mass-squared difference is.

↘ Derive: why oscillation requires mass

Suppose the three neutrino flavours ν_e, ν_μ, ν_τ are not mass eigenstates, but linear combinations of three mass eigenstates ν₁, ν₂, ν₃ with masses m₁, m₂, m₃. The mixing is parametrised by the PMNS matrix U:

|ν_α⟩ = Σ_i U_αi |ν_i⟩

Each mass eigenstate, in flight, picks up a phase governed by its energy. For a neutrino with momentum p, in the ultra-relativistic limit E_i ≈ p + m_i² / (2p) ≈ E + m_i² / (2E):

|ν_i(t)⟩ = e^(−i E_i t) |ν_i(0)⟩

The probability that a neutrino produced as flavour α is detected as flavour β after travelling a distance L ≈ t (in natural units) is:

P(ν_α → ν_β) = |⟨ν_β|ν_α(t)⟩|² = |Σ_i U_βi* U_αi e^(−i m_i² L / 2E)|²

Expand the modulus for two flavours, mixing angle θ:

P(ν_α → ν_β) = sin²(2θ) · sin²(Δm² · L / 4E)

with Δm² = m_j² − m_i² and L/E in suitable units. The key point: the formula is identically zero if Δm² = 0. Massless neutrinos cannot oscillate, no matter how much mixing you put in. Conversely, an observed oscillation directly measures Δm². Atmospheric data give Δm²_atm ≈ 2.5 × 10⁻³ eV²; solar data give Δm²_sol ≈ 7.5 × 10⁻⁵ eV². The first oscillation maximum, where sin²(Δm² L / 4E) = 1, sits at L/E = π / (Δm²) in eV²·km/GeV units, which works out to roughly 500 km/GeV for the atmospheric splitting and 30000 km/GeV for the solar one. That is the structure of the figure above.

Three flavours have three angles (θ₁₂, θ₂₃, θ₁₃), two mass splittings, and one CP-violating phase δ_CP. The angles are now all measured (and large: θ₂₃ is near 45°, much bigger than any analogous angle in the quark sector). The phase δ_CP is the active frontier: T2K and NOvA both prefer values near −π/2, hinting that neutrinos may violate CP differently from antineutrinos. If true, leptogenesis (neutrinos as the asymmetric agent that left matter outnumbering antimatter after the Big Bang) becomes a viable origin story for everything you can see.

We now know neutrinos have mass, and we know roughly how much. What we do not know is why they have so little. The Higgs mechanism, the trick that gives every other Standard Model particle its mass, works by coupling left-handed particles to the Higgs field, with the coupling strength setting the mass. The top quark couples strongly (mass 172 GeV); the electron weakly (mass 0.000511 GeV); the up quark weakly (mass about 0.002 GeV). The pattern is roughly that masses are set by Yukawa couplings of order unity for the heaviest particles, falling away geometrically. If neutrinos got their mass the same way, you would expect a Yukawa coupling of, very roughly, 10⁻¹² to produce a mass of 0.05 eV. That is fourteen orders of magnitude smaller than the top’s coupling, and twelve orders smaller than the electron’s. It does not feel right.

The most popular explanation is called the seesaw mechanism, and it works like this. Suppose there exists, in addition to the three left-handed neutrinos we observe, a set of very heavy right-handed neutrinos N with mass M near the grand-unification scale, perhaps 10¹⁴ GeV. Both light and heavy neutrinos couple to the Higgs with ordinary, order-one strength. Diagonalising the resulting 2 × 2 mass matrix gives two eigenvalues. One is heavy, very close to M itself. The other is tiny, of order m_D² / M, where m_D is the ordinary Dirac mass scale set by Higgs coupling (call it 100 GeV). Plug in numbers: (100 GeV)² / (10¹⁴ GeV) = 10⁻¹⁰ GeV = 0.1 eV. The arithmetic falls out exactly where the cosmological mass bound puts the heaviest neutrino. The lightness of the observed neutrinos becomes a measure of the heaviness of an invisible partner.

The seesaw is elegant, and the heavy right-handed neutrinos have an extra job they can do for free: by decaying out of equilibrium in the early universe, with a CP-violating phase like δ_CP, they can leave a tiny preference for matter over antimatter. The mechanism is called leptogenesis, and if it is right, the reason the observable universe is not just photons is that neutrinos, ninety-six years after Pauli invented them on a Zürich evening, broke a quantum symmetry while the universe was a microsecond old. It is the kind of explanation that physicists call “beautiful” and engineers call “untestable until we can build a 10¹⁴ GeV accelerator.” Possibly forever.

The neutrino remains, ninety-six years after Pauli’s letter, the strangest particle in the Standard Model. Every other particle’s mass can be expressed cleanly in terms of a Yukawa coupling times the Higgs vacuum expectation value, and the couplings, though they span six orders of magnitude, are at least all of the same logical type. The neutrino refuses. Either its Yukawa is twelve orders of magnitude smaller than the electron’s (which begs an explanation), or its mass comes from a different mechanism entirely (the seesaw, involving heavy right-handed partners we have not seen and probably never will see directly), or it is its own antiparticle (Majorana, in which case neutrinoless double-beta decay should occur and the current generation of experiments are looking hard). Each possibility implies a different completion of physics beyond the Standard Model. Each is a different bet about what the universe is made of at scales we cannot yet reach.

The deepest implication may be the one Pauli would have appreciated. He invented the neutrino to save energy conservation in beta decay. The neutrino, in repayment, may turn out to have saved the universe from being empty. If the seesaw is right and leptogenesis is real, then the same heavy right-handed neutrinos whose existence we infer from the lightness of the visible ones, acting in the first microsecond of cosmic history, produced the tiny matter-antimatter asymmetry without which the universe would have annihilated itself into pure photons before the first nucleus formed. Every atom in your body, every star you can see, every page of physics ever written, would all be owed to a CP-violating phase in the mixing matrix of three of the slipperiest, lightest, most stubbornly invisible particles ever postulated. Pauli sent his apology letter to a meeting he could not attend. The universe sent back its existence.

There is one piece of unfinished business. The Standard Model proper still includes only the three light neutrinos and treats their masses as a set of free parameters slotted in by hand. The deeper picture, the one in which mass itself is explained rather than enumerated, is the one the next chapter takes up. To get there we leave neutrinos and turn to the strong force, where the constituents of every proton and neutron in your body are stuck together by a binding energy so violent it is responsible for ninety-nine percent of the mass of ordinary matter, and where the rule is not lightness but confinement.

§

The neutrino problem was a problem of too little mass, hiding in plain sight. The next problem is the opposite: too much mass, generated by binding energy in a force so strong that the carriers of the force are themselves trapped inside the particles they bind. Quarks are never found alone. Why?

next chapter → Color confinement