Borexino's CNO Neutrinos: Catching the Sun's Catalytic Cycle

The Sun derives its energy from nuclear fusion in its core, where temperatures of 15 million Kelvin and densities of 150 g/cm³ allow protons to overcome their mutual electrostatic repulsion. Of the two main fusion pathways — the proton-proton (pp) chain and the carbon-nitrogen-oxygen (CNO) cycle — the pp chain accounts for approximately 99% of solar energy production at the Sun’s metallicity. The CNO cycle contributes the remaining ~1%, but its rate depends on the abundance of C, N, and O in the core, making it a sensitive probe of stellar composition.

For decades, neutrino-based confirmation of the Sun’s CNO cycle was a frontier goal. The pp-chain neutrinos had been measured by multiple experiments — Homestake (chlorine), GALLEX/SAGE (gallium), Super-Kamiokande and SNO (water), Borexino itself (specific spectral lines). The CNO neutrinos, however, are an order of magnitude rarer, with energies spanning a continuous spectrum that overlaps with both the pep neutrinos (from the pp chain) and natural radioactive backgrounds. They had never been directly detected.

In June 2020, the Borexino collaboration announced that the long-awaited CNO measurement had been achieved. After fifteen years of running with one of the lowest natural-radioactivity scintillators ever built, Borexino reported approximately 5σ evidence for solar CNO neutrinos. The CNO cycle — operating in our Sun, fusing protons into helium with C-N-O catalysts — had been directly confirmed.

This post is about how Borexino made the measurement, why it was so difficult, and what the result implies for the standard solar model.

The CNO cycle in detail

In the CNO bicycle, four protons fuse into one helium-4 nucleus through a chain of reactions catalyzed by C, N, and O nuclei:

${}^{12}\text{C}+p{\to }^{13}\text{N}+\gamma$ ${}^{13}\text{N}{\to }^{13}\text{C}+e^{+}+{\nu }_e(Q\approx 1.2\text{ MeV})$ ${}^{13}\text{C}+p{\to }^{14}\text{N}+\gamma$ ${}^{14}\text{N}+p{\to }^{15}\text{O}+\gamma$ ${}^{15}\text{O}{\to }^{15}\text{N}+e^{+}+{\nu }_e(Q\approx 1.7\text{ MeV})$ ${}^{15}\text{N}+p{\to }^{12}\text{C}{+}^4\text{He}$

The carbon-12 catalyst is recovered at the end of the cycle. The two intermediate beta decays (^{13}N and ^{15}O) emit electron neutrinos with continuous spectra extending up to 1.2 and 1.7 MeV respectively. A third minor branch produces ^{17}F neutrinos with even higher endpoint, but these are about 100 times rarer and inaccessible at Borexino’s sensitivity.

The rate of the CNO cycle depends on:

• The temperature of the stellar core
• The proton density
• The abundance of C, N, O nuclei (the catalysts)

Since the temperature and density are well-constrained by helioseismology and the standard solar model, the dominant uncertainty in the predicted CNO neutrino flux is the elemental abundance — the metallicity of the solar core.

The metallicity problem

The Sun’s metallicity is fundamental to stellar physics, but its precise value is uncertain.

Two principal methods give incompatible answers:

Spectroscopic measurements of the solar surface atmosphere, using high-resolution spectral lines, give a relatively low metallicity (Asplund et al. 2009: $Z=0.0134$, where $Z$ is the mass fraction of elements heavier than helium).

Helioseismic measurements of the solar interior, using sound-speed measurements via solar oscillation modes, give a relatively high metallicity ($Z\approx 0.018$).

The two values differ by approximately 30%. The discrepancy has persisted for over a decade and is one of the unresolved questions in stellar physics. It has consequences for stellar evolution models, for the calibration of stellar ages, and for the inferred composition of the early universe.

A precise CNO neutrino flux measurement could resolve the metallicity question. The flux is approximately a factor of 1.5 higher in the high-metallicity case than in the low-metallicity case — a substantial difference at the precision Borexino was aiming for.

Borexino’s design

Borexino — the Borexino solar neutrino experiment in Italy’s Gran Sasso laboratory — was designed specifically for low-energy solar-neutrino measurements. The detector is a 280-tonne sphere of ultra-pure liquid scintillator (PC-PPO mixture), surrounded by a 2,200-tonne water-Cherenkov muon veto, with the entire apparatus in a 1,400-metre-deep underground laboratory shielded from cosmic-ray backgrounds.

The key feature is radiopurity. The scintillator is the cleanest substance ever produced for neutrino physics: ${}^{238}$U content of approximately $10^{-19}$ g/g, ${}^{232}$Th of approximately $10^{-19}$ g/g, ${}^{14}$C of approximately $10^{-18}$, and other contaminants similarly low. These are 8-10 orders of magnitude lower than commercial-grade liquid scintillators.

The detection signal: solar neutrinos elastically scatter off electrons in the scintillator. The recoiling electron deposits energy that produces scintillation light. The light is detected by approximately 2,000 photomultiplier tubes surrounding the scintillator volume.

For low-energy solar neutrinos (below about 1 MeV), the dominant backgrounds are:

• ${}^{14}$C beta decay (endpoint 156 keV) — only relevant below 200 keV
• ${}^{210}$Bi beta decay (endpoint 1.16 MeV) — directly in the CNO energy range
• Solar pep neutrinos (monoenergetic at 1.44 MeV) — also in the CNO energy range
• Cosmic-muon-induced cosmogenic isotopes — controlled by the depth of the laboratory

The CNO measurement requires subtraction of ${}^{210}$Bi and pep contributions from the observed energy spectrum.

The 210Bi challenge

${}^{210}$Bi is the most insidious background for CNO. It is part of the natural ${}^{238}$U decay chain, and its energy spectrum is similar to the CNO spectrum. Borexino’s scintillator initially had a ${}^{210}$Bi rate of approximately 75 counts per day per 100 tonnes — already extraordinarily low, but not low enough.

The collaboration’s solution: monitor the scintillator’s decay over time and exploit the fact that ${}^{210}$Po decays from ${}^{210}$Bi with a 138-day half-life. By measuring the ${}^{210}$Po rate (which has a clean signature) and assuming secular equilibrium, the ${}^{210}$Bi rate can be inferred. Combined with stable thermal stratification (the scintillator is held at slightly different temperatures at different depths to prevent convection), the ${}^{210}$Bi distribution can be precisely characterized and subtracted.

Borexino’s analysis also used deep statistical techniques — including improved understanding of the scintillator’s behavior under thermal control — to constrain the ${}^{210}$Bi background to high precision. Over the years 2015-2020, this enabled the CNO measurement.

The pep neutrino constraint

The other major background — solar pep neutrinos — overlaps spectrally with CNO. Borexino constrained the pep flux through an independent measurement: pep neutrinos are monoenergetic at 1.44 MeV, producing a sharp Compton-edge feature in the energy spectrum. By combining the spectroscopic identification of pep with luminosity-conservation arguments (the solar luminosity is precisely known), the pep rate is constrained to approximately 5% precision.

With both ${}^{210}$Bi and pep characterised, the residual signal in the 0.7-1.5 MeV energy range is the CNO contribution.

The 2020 measurement

The original Borexino CNO measurement, published in Nature 587, 577 (2020), reported the CNO flux as: ${\Phi }_{CNO}=(7.0\pm 1.5\text{stat}\pm 1.0\text{syst})\times 10^8{\text{ cm}}^{-2}{\text{s}}^{-1}$

This is consistent with both the high-metallicity and low-metallicity predictions:

• High-metallicity prediction: $\sim 4.9\times 10^8$ cm⁻²s⁻¹ (CNO flux at oscillation-corrected production rate)
• Low-metallicity prediction: $\sim 3.5\times 10^8$ cm⁻²s⁻¹

The Borexino result is closer to the high-metallicity prediction but with sufficient uncertainty that neither metallicity model is excluded. The 5σ statistical significance establishes that CNO neutrinos exist as a measurable component of the solar flux.

The 2022 update (Bagdasarian et al., Physical Review Letters 129, 252701) tightened the result with additional Phase II data: ${\Phi }_{CNO}=(6.7\pm 1.2)\times 10^8{\text{ cm}}^{-2}{\text{s}}^{-1}$

Same conclusion: high-metallicity preference but not exclusion of low-metallicity.

The metallicity question, partially resolved

With the 2022 update, Borexino’s result slightly favors the high-metallicity solar model. The high-metallicity value is consistent with the helioseismic measurements; the low-metallicity is the spectroscopic result. So Borexino’s CNO measurement provides a third independent probe of the solar metallicity that, within its precision, supports the helioseismic side.

This is significant because the spectroscopic measurements (which give low metallicity) are based on absorption-line analyses of the solar photosphere — a region with complex 3D atmospheric models and uncertain departures from local thermodynamic equilibrium. The analyses depend on assumptions that have been updated multiple times in the past decade. The CNO neutrino measurement is independent of these atmospheric modeling uncertainties.

The current consensus, as of 2026, is that the metallicity question is leaning toward the high-metallicity (helioseismic) solution, with the spectroscopic methods needing further refinement. Future CNO measurements will tighten this conclusion.

What’s next

Borexino was decommissioned in 2021. Its scintillator was preserved and is being used for various smaller-scale experiments. Most importantly, its expertise has flowed into successor projects.

SNO+ at Sudbury, Canada — using the original SNO cavity now refilled with liquid scintillator — is one of the leading candidates for an improved CNO measurement. SNO+‘s scintillator uses LAB (linear alkylbenzene) plus PPO, similar to Borexino’s recipe but with potentially higher light yield. With a 780-tonne fiducial mass (compared to Borexino’s 75 tonnes effective for solar neutrinos), SNO+ has the statistical power to push the CNO measurement to <10% precision. Operations are underway.

Theia — a hybrid water-and-scintillator detector concept under development — would combine the rapid event reconstruction of water Cherenkov with the scintillation light yield of organic liquid scintillator. Such a detector at multi-kiloton scale could measure CNO neutrinos at <5% precision.

By 2030-2035, the metallicity question should be definitively resolved through CNO neutrino measurements at next-generation experiments. The current Borexino result is the foundation; the precision will improve substantially.

A different kind of milestone

The Borexino CNO measurement is unusual in modern neutrino physics. It is not a discovery of new physics, nor a confirmation of an oscillation parameter, nor a search for beyond-Standard-Model effects. It is a direct, in-situ measurement of a fundamental stellar process — confirming that the Sun does, indeed, fuse hydrogen via the CNO catalytic cycle, exactly as stellar models predict.

This kind of result is rare because the CNO contribution to solar luminosity is small (~1%), the neutrino flux is correspondingly weak, and the relevant energy range is buried in backgrounds. Only a detector with Borexino’s combination of radiopurity, statistical power, and analysis sophistication could have achieved it.

The result is also a measurement of stellar interior physics through a particle-physics technique. The rate at which the Sun’s catalytic cycle operates depends on its core composition. By catching the resulting neutrinos, Borexino has provided astrophysics with information about a region that no other experiment can probe. In this sense, the measurement bridges particle physics and stellar astrophysics in a way that few other observations do.

The CNO cycle, predicted by Hans Bethe in 1939 and integrated into stellar models throughout the 20th century, is now experimentally confirmed in our nearest star. Eighty years of theoretical and observational work, capped by a single neutrino-counting result.

The Sun fuses hydrogen via two pathways. Both have now been directly measured. The standard solar model — the foundation of stellar astrophysics — is confirmed to be substantially correct. The remaining uncertainties are in the details, and those details, in time, will be measured.