← quantum · contents

§ Phase xv · Stellar Quanta · Chapter 01

Stellar fusion

The sun's core, at fifteen million degrees, is far too cold by any classical reckoning for two protons ever to meet. Each carries a positive charge, each repels the other with a Coulomb wall roughly a thousand times higher than the kinetic energy it can muster, and yet the sun has been quietly fusing four million tonnes of mass into light every second for four and a half billion years. The reconciliation is the same exponential leak that explains alpha decay, applied in reverse: protons tunnel inward, not outward.

In 1920, Arthur Eddington stood at the British Association for the Advancement of Science meeting in Cardiff and gave a presidential address that would set the agenda for the next thirty years of astrophysics. He had been reading Francis Aston’s mass-spectrograph measurements, which showed that a helium nucleus weighs slightly less than four hydrogen nuclei. The mass deficit, about seven parts in a thousand, was small. Multiplied by Einstein’s c², it was enormous: enough energy, Eddington calculated, to keep a star like the sun shining for billions of years, far longer than any chemical or gravitational source could manage. “If, indeed, the sub-atomic energy in the stars is being freely used to maintain their great furnaces,” he said, “it seems to bring a little nearer to fulfilment our dream of controlling this latent power for the well-being of the human race, or for its suicide.” That last clause was uncomfortably prescient.

The mechanism Eddington proposed was simple enough to state and impossible to defend. Four hydrogen nuclei somehow merged into one helium nucleus inside the star. Two of the protons turned into neutrons along the way, expelling the surplus charge as positrons. The mass that disappeared became the light and heat that the star radiated. Eddington could see the bookkeeping, and the bookkeeping was perfect. What he could not explain was the mechanism. Two protons at the temperature of the solar core have a kinetic energy of about a kilo-electron-volt, while the electrostatic wall between them rises to nearly an MeV. The collision energy is short of the barrier by a factor of a thousand. Classically, the protons would bounce off each other long before any nuclear force could grip them. The reaction simply could not happen.

Eddington was unmoved. To the audience that called his theory impossible, he offered one of the most-quoted retorts in twentieth-century science: “We do not argue with the critic who urges that the stars are not hot enough for this process; we tell him to go and find a hotter place.” Privately he was less confident. He spent the next decade waiting for someone to find a way out, and for several years it looked as if nobody would.

The escape from Eddington’s impasse came, like so much of 1920s physics, from quantum mechanics. The crucial paper was published in 1928 by a freshly-minted Soviet doctorate at Göttingen, the same paper that solved alpha decay. George Gamow had been thinking about radioactive nuclei, but the inverse problem followed immediately. If an alpha particle could leak out of a uranium nucleus by tunneling through its own Coulomb wall, then a proton in a star’s core could tunnel in through another proton’s wall. The barrier and the wavefunction did not care which direction the arrow of probability pointed.

Gamow’s tunneling factor for two charged particles meeting head-on takes a beautiful form. Write v for their relative speed and η for the dimensionless ratio

η = Z₁ Z₂ e² / (ℏ v)

where Z₁ and Z₂ are the charges (one each, in the proton-proton case) and e² here absorbs the usual factor of one over four pi epsilon zero. The probability that the two particles tunnel into nuclear contact, instead of bouncing off the Coulomb wall, is

P ~ exp(-2π η)

which is now universally called the Gamow factor. The exponent is enormous when v is small. At solar core temperatures, where the typical proton speed is about a thousand kilometres per second, 2π η is around twenty-two, so P is roughly exp(-22) ≈ 3 × 10⁻¹⁰. One in three billion proton encounters tunnels. That seems hopeless until you remember that the sun’s core has about 10⁵⁶ protons each rattling against its neighbours roughly 10¹⁸ times a second. Even one in three billion, multiplied by 10⁷⁴ collisions per second, gives a vast reaction rate. The exponential is doing the violence; the geometry of the star is doing the patience.

Atkinson and Houtermans applied Gamow’s formula to stars within a year of the original paper, in a 1929 article titled “How can one cook helium nucleus in a potential cauldron?”, and concluded that thermonuclear fusion was at least possible. They did not yet have a specific reaction in mind. The energy was there, the tunneling factor was small but not zero, the temperatures matched up. The detailed chemistry of how four protons combine into one helium nucleus remained open for another decade.

A decade later, the missing chemistry arrived in a single sustained burst of work by Hans Bethe. Bethe had fled Germany in 1933, landed at Cornell by way of Manchester and Bristol, and quickly become the leading theorist of nuclear reactions in the United States. In the spring of 1938, his friend George Gamow organised a small conference in Washington on the problem of stellar energy. Bethe attended reluctantly. He was a nuclear physicist, not an astronomer, and he later said he had no interest in stars before the meeting started. By the end of the conference he had the answer. He went home, worked steadily for six months, and in early 1939 submitted to Physical Review a 31-page paper titled “Energy Production in Stars.” It was the most influential paper in stellar astrophysics ever written.

Bethe’s paper laid out two distinct fusion chains, both of which take four protons in and put one helium-4 nucleus out, releasing 26.7 MeV of binding energy. The first, which had been sketched in pieces by Critchfield and others, became the proton-proton chain. It begins with the slowest step in the entire sun: two protons collide, tunnel through their Coulomb wall, and one of them undergoes beta decay during the brief moment they are touching. The product is a deuteron, a positron, and a neutrino. The rate of this step is so low that the average proton in the sun waits about a billion years for it to happen. The remaining chemistry runs much faster. The deuteron grabs another proton within a second to form helium-3, and then two helium-3 nuclei merge into helium-4 with two protons spat back out. Net result: four protons in, one helium-4 nucleus out, two positrons, two neutrinos, and 26.7 MeV of energy, about 0.6 MeV of which slips out as neutrinos and is lost to the star.

That sequence is called pp-I, and in the sun it accounts for about 86 percent of the energy. Two sister branches, pp-II and pp-III, share the same starting steps but take a detour through beryllium-7 and (for pp-III) boron-8. They are rarer in the sun and produce neutrinos at higher energies, which turned out to matter for the experimental story decades later.

Bethe’s second chain, the CNO cycle, used carbon, nitrogen, and oxygen nuclei as catalysts. A carbon-12 nucleus captures a proton, undergoes a series of beta decays and proton captures, and emerges as carbon-12 again with a helium-4 nucleus subtracted from four protons. The catalyst is not consumed. Bethe noted that the CNO rate is far more temperature-sensitive than the pp-chain, scaling roughly as the temperature to the seventeenth power versus the fourth. In the sun, where the core sits at about fifteen million Kelvin, the pp-chain wins by a wide margin. But in any star with a core hotter than about seventeen million Kelvin (very roughly, more massive than 1.3 solar masses), the CNO cycle takes over and dominates. Bethe had explained, in one paper, why low-mass stars and high-mass stars look so different.

There is one detail of the pp-chain worth dwelling on, because it embarrasses textbooks. The very first step, p + p → d + e⁺ + ν, is not a strong-interaction reaction. It cannot be. Two protons forming a deuteron requires one of them to change into a neutron, and that requires the weak force. The first step in the chain is therefore a weak-interaction event happening during a quantum-mechanical tunneling encounter. Both the Coulomb tunneling and the weak decay must coincide. The cross-section is preposterously small, around 10⁻⁴⁷ square centimetres, smaller than any reaction that has ever been measured in a laboratory. Nobody has ever observed a proton-proton fusion event directly. The reaction rate is inferred entirely from theory plus the solar neutrino flux, and the theoretical calculation is one of the few places in physics where the standard model has been tested in a regime no accelerator can ever reach.

Bethe’s paper sat for nearly thirty years before the Nobel committee got around to it. In 1967 he received the prize “for his contributions to the theory of nuclear reactions, especially his discoveries concerning the energy production in stars.” He had spent most of the intervening decades on other things. The wartime years went into the Manhattan Project, where he ran the theoretical division at Los Alamos. The postwar decades went into shock waves, supernovae, neutron-star structure, quantum electrodynamics, and the political arguments over arms control. Bethe lived to 98. He never lost his loyalty to the pp-chain.

The pp-chain produced a side story that ran for a third of a century and ended in a Nobel prize of its own. Bethe’s chemistry predicts that each fusion event releases two neutrinos, almost all of them with energies between zero and about a few MeV. The neutrinos pass through the sun with negligible scattering and reach Earth eight minutes later. Sixty billion of them cross every square centimetre of your body every second, and almost none of them interact. In the early 1960s, Ray Davis, working in a chemistry lab at Brookhaven, set out to catch some of them anyway.

Davis built a tank holding 615 tonnes of perchloroethylene, the common dry-cleaning fluid, and installed it 4,850 feet underground in the Homestake gold mine in South Dakota. The depth was the trick: the rock filtered out cosmic-ray backgrounds so that any chlorine-37 nucleus turning into argon-37 was very probably caused by a neutrino. The reaction has a threshold of about 0.81 MeV, which meant that Davis was sensitive only to the higher-energy neutrinos, mostly from the pp-III branch through boron-8 decay. The boron-8 flux is tiny, perhaps one part in ten thousand of the total solar neutrino output, but it is fierce enough to be detectable. The expected rate was about one argon atom per day. Davis counted argon atoms.

What he found, year after year from 1968 onward, was about a third of what theory predicted. The discrepancy was at first dismissed as either an experimental error on Davis’s part or a modelling error in John Bahcall’s elaborate standard-solar-model calculations of the boron-8 flux. Both possibilities were tested, exhaustively. Neither resolved the deficit. Through the 1970s and 1980s the solar neutrino problem became a chronic embarrassment to nuclear astrophysics. The conventional view was that something was wrong with the sun: maybe the core was cooler, maybe the metallicity was off, maybe the convection zone was deeper than thought. Cleaner experiments at Kamiokande in Japan and SAGE and GALLEX in Italy and Russia confirmed Davis’s deficit and pushed it to lower energies. The pattern of the deficit was peculiar, however. It did not look like a small adjustment to the sun. It looked like a fraction of the neutrinos were vanishing somewhere between their birthplace and the detector.

The resolution came in 2001. The Sudbury Neutrino Observatory, an underground tank of 1,000 tonnes of heavy water in a nickel mine in Ontario, was sensitive to all three neutrino flavours rather than electron neutrinos alone. SNO measured both the electron-neutrino flux from the sun and the total flux of all three flavours. The electron-neutrino number was one-third of the prediction, matching Davis. The all-flavour number was the full prediction. The missing two-thirds had not vanished. They had turned into muon and tau neutrinos on the way out of the sun, through a quantum-mechanical oscillation that requires neutrinos to have a tiny but nonzero mass. SNO collected its 2015 Nobel prize for this; Davis collected his in 2002. Bethe’s chemistry, which had predicted a flux Davis could not see, turned out to be exactly right. The flux was there. It was the neutrinos that had been hiding.

ε_ij = (n_i n_j) / (1 + δ_ij) · ⟨σv⟩ · Q

⟨σv⟩ ∝ T^(-2/3) · exp(-3 (b² / (4 kT))^(1/3))

There is a final twist that Hoyle and Fowler added in the 1950s. Bethe’s chains take you from hydrogen to helium, and you can run them forward through helium burning into carbon, oxygen, and so on. Fred Hoyle realised that the helium-to-carbon step needed a quantum loophole of its own. Three helium-4 nuclei must combine into one carbon-12, but the simultaneous collision of three particles is preposterously rare. Hoyle predicted that there had to be an unknown excited state of carbon-12, at exactly the right energy to be reached by two helium-4 nuclei combining (into beryllium-8, which lives only 10⁻¹⁶ seconds) and then capturing a third helium-4 before the beryllium falls apart. Without that resonance, the universe contains no carbon, no oxygen, no biology. Hoyle predicted the state on anthropic grounds and persuaded William Fowler to look for it. Fowler’s group found the state exactly where Hoyle said it would be. It is now called the Hoyle state, and it is the reason you exist.

The universe, as Hoyle put it years later, “looks like a put-up job.” From a quantum-mechanical perspective, every step in the journey from primordial hydrogen to a carbon-based biosphere is a tunneling event with an exponentially small probability, multiplied against a stellar reservoir large enough to make the exponential survivable. The sun is one tunneling clock running at one frequency. Eddington’s “find a hotter place” critic was wrong in the spirit and right in the letter. There is a hotter place: it is just a barrier away.