CNO Cycle

Every main-sequence star runs on the same net transaction — four protons fused into one helium-4 nucleus, worth about $26.73\,\mathrm{MeV}$ — but nature offers two assembly lines for it. In cool, low-mass cores the proton–proton chain does the work directly. In hotter cores the carbon–nitrogen–oxygen (CNO) cycle takes over, and it never consumes its C, N, and O nuclei at all: those heavier nuclei act as catalysts, temporarily hosting captured protons before ejecting an alpha and handing back the original nucleus, ready to run the loop again. Why does a hotter star switch assembly lines? The Coulomb barrier. Fusing a proton onto a nucleus of charge $Z=6$ – $8$ means tunnelling through a barrier six to eight times taller than the one between two protons, and the Gamow-peak analysis turns that taller barrier into a rate that climbs far more steeply with temperature. The consequence is dramatic: the CNO cycle supplies barely a percent of the Sun's luminosity at $1.5\times10^{7}\,\mathrm{K}$ , yet becomes overwhelmingly dominant by $2$ – $3\times10^{7}\,\mathrm{K}$ , powering every star above about $1.3\,M_\odot$ . That steepness hides a slow step. The reaction $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ is the bottleneck of the whole loop, so catalytic material queues up as nitrogen — the fingerprint of CNO processing seen at the surfaces of evolved stars. And because the burning is so centrally concentrated, massive stars grow convective cores that reshape the upper main sequence. This article assembles all of that from the Gamow peak, verifies the celebrated $T^{18}$ scaling by hand, and follows the physics out to solar neutrinos and classical novae.

Key takeaways

The CNO cycle fuses four protons into one helium-4 using carbon, nitrogen, and oxygen as catalysts: the seed nuclei are destroyed and regenerated every loop, so only trace CNO abundances are needed.
Because it fuses protons onto nuclei of charge $Z=6,7,8$ , the Coulomb barrier is far higher than for the pp chain, making the rate extraordinarily temperature-sensitive — $\varepsilon_{\mathrm{CNO}}\propto T^{\sim 18}$ near $2\times10^{7}\,\mathrm{K}$ .
That steepness lets it overtake the pp chain above a core temperature of about $1.7\times10^{7}\,\mathrm{K}$ , i.e. in stars more massive than roughly $1.3\,M_\odot$ .
The slowest step is proton capture on $^{14}\mathrm{N}$ , so in equilibrium the catalysts pile up as nitrogen — carbon is driven into nitrogen and $^{12}\mathrm{C}/^{13}\mathrm{C}$ falls toward $\approx 3$ – $4$ .
The net release is the same $\sim 26.73\,\mathrm{MeV}$ per He-4 as the pp chain, but a larger fraction ( $\sim 6\%$ ) escapes as neutrinos, and the strong $T$ -dependence builds convective cores in massive stars.
Key idea: C, N, O act as catalysts, not fuel — regenerated each loop so their total abundance is conserved while four protons are net-fused into one He-4; the rate-limiting $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ step plus the high Coulomb barrier give the cycle its huge temperature sensitivity and drive C into N.

The core insight

C, N, O act as catalysts, not fuel: they are regenerated each loop, so their total abundance is conserved while four protons are net-fused into one He-4. The rate-limiting $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ step and the high Coulomb barrier give the cycle its huge temperature sensitivity and drive C into N.

Where this lives today

The CNO cycle sits at several live research fronts. The sharpest recent result is Borexino's 2020 detection of solar CNO neutrinos — the first direct evidence the cycle runs in the Sun. Though it makes only about 1% of the Sun's power, the flux of $^{13}\mathrm{N}$ , $^{15}\mathrm{O}$ , and $^{17}\mathrm{F}$ neutrinos is proportional to the carbon-plus-nitrogen abundance in the core, making it a rare in-situ probe of solar metallicity — the quantity at the heart of the "solar abundance problem." On the nuclear side the bottleneck rate $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ is still being refined: underground accelerators such as LUNA at Gran Sasso measure its low-energy cross section, and revisions there shift globular-cluster ages by hundreds of millions of years. The cycle is also the dominant source of $^{14}\mathrm{N}$ in the universe, and the hot CNO cycle governs the energetics of classical novae and type-I X-ray bursts — its beta-limited breakout reactions a target for radioactive-beam facilities.

The physical picture

Picture the C, N, and O nuclei as reusable buckets on a conveyor. A proton drops into a bucket ( $(p,\gamma)$ capture), the overfilled bucket settles by beta-plus decay, collects more protons, and eventually spits out a helium nucleus, returning to its original form. Four protons went in over the loop, one alpha came out, and the bucket is unchanged. That is what "catalyst" means quantitatively: the number of C+N+O nuclei is conserved, and a trace abundance — a metallicity of a fraction of a percent — suffices to process the star's whole hydrogen supply, because each catalyst is reused millions of times.

Figure

The CN cycle runs clockwise from ¹²C through (p,γ) and (p,α) captures around a closed loop that regenerates the catalyst while netting 4 ¹H → ⁴He + 2e⁺ + 2νₑ + γ with Q ≈ 26.73 MeV; the slowest step ¹⁴N(p,γ)¹⁵O is the bottleneck where nitrogen accumulates

The cycle is fundamentally a bottleneck problem, and that is where its chemistry comes from. Of the six reactions, $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ has by far the smallest cross section, so material queues up behind it like cars behind the slowest toll booth: in steady state each isotope's abundance is inversely proportional to its destruction rate, so the slow-reacting nitrogen becomes the most abundant catalyst, and nearly all the carbon a star was born with ends up as $^{14}\mathrm{N}$ . The extreme temperature sensitivity follows from the Gamow peak: a higher nuclear charge means a taller, thicker Coulomb barrier, so the fraction of pairs able to tunnel rises steeply with $T$ . The second figure shows the practical consequence.

Figure

Energy generation versus core temperature on log–log axes: the pp chain rises as ε ∝ T⁴ while the CNO cycle rises as ε ∝ T¹⁸, so the steep CNO curve overtakes it at a crossover near T ≈ 1.7×10⁷ K — the central temperature of a ≈1.3 M⊙ star, above which CNO burning dominates

The core equations

Everything starts from one accounting statement: the net transmutation is identical to the pp chain — four protons into one alpha — and the catalysts never appear because they are recovered. The energy $Q$ is the mass deficit, released as gamma rays, positron annihilation, and neutrinos.

4\,{}^{1}\mathrm{H}\;\longrightarrow\;{}^{4}\mathrm{He}+2e^{+}+2\nu_e+\gamma,\qquad Q\approx 26.73\,\mathrm{MeV} \qquad (1)

where: $Q$ — total energy released per He-4 formed, including the two positron annihilations; $\nu_e$ — electron neutrino, of which two are emitted, carrying away a larger share than in the pp chain. Reading it: the CNO nuclei are absent from the net equation precisely because they are catalysts. The dominant branch is the CN cycle (CNO-I), a closed loop of six reactions on $^{12}\mathrm{C}$ : two beta-plus decays and four proton captures — three radiative $(p,\gamma)$ and one $(p,\alpha)$ that closes the loop.

\begin{aligned}{}^{12}\mathrm{C}(p,\gamma){}^{13}\mathrm{N}&\;\;{}^{13}\mathrm{N}\to{}^{13}\mathrm{C}+e^{+}+\nu_e\\{}^{13}\mathrm{C}(p,\gamma){}^{14}\mathrm{N}&\;\;{}^{14}\mathrm{N}(p,\gamma){}^{15}\mathrm{O}\\{}^{15}\mathrm{O}\to{}^{15}\mathrm{N}+e^{+}+\nu_e&\;\;{}^{15}\mathrm{N}(p,\alpha){}^{12}\mathrm{C}\end{aligned} \qquad (2)

where: $(p,\gamma)$ — proton capture emitting a gamma ray; $(p,\alpha)$ — proton capture ejecting an alpha, closing the loop. Reading it: the mass number climbs by ones from 12 to 15, drops back to 12 when the alpha is released, and the two beta decays supply the two neutrinos in equation (1). A small fraction of the flux leaks into an oxygen branch (the ON or CNO-II cycle) when $^{15}\mathrm{N}$ captures a proton radiatively to make $^{16}\mathrm{O}$ instead of ejecting an alpha; the material eventually re-enters the CN loop as $^{14}\mathrm{N}$ .

{}^{15}\mathrm{N}(p,\gamma){}^{16}\mathrm{O}\;\to\;{}^{16}\mathrm{O}(p,\gamma){}^{17}\mathrm{F}\;\to\;{}^{17}\mathrm{O}(p,\alpha){}^{14}\mathrm{N} \qquad (3)

where: $^{16}\mathrm{O}$ — enters when $^{15}\mathrm{N}(p,\gamma)$ wins over $^{15}\mathrm{N}(p,\alpha)$ . Reading it: this leak occurs only about once per $10^{3}$ – $10^{4}$ CN cycles, but over stellar lifetimes it links the star's oxygen into the same catalytic pool. The energy generation rate is what the stellar-structure equations need — linear in density and total catalyst abundance, carrying the enormous effective temperature exponent that is the cycle's signature.

\varepsilon_{\mathrm{CNO}}\approx \varepsilon_{0}\,\rho\,X\,X_{\mathrm{CNO}}\,T^{\nu},\qquad \nu\equiv\frac{\mathrm{d}\ln\varepsilon}{\mathrm{d}\ln T}\approx 18 \qquad (4)

where: $\varepsilon_{\mathrm{CNO}}$ — energy generated per unit mass per unit time ( $\mathrm{erg\,g^{-1}\,s^{-1}}$ ); $\rho$ — mass density; $X$ — hydrogen mass fraction; $X_{\mathrm{CNO}}$ — combined mass fraction of C, N, O; $\nu$ — local temperature exponent, $\approx 18$ near $2\times10^{7}\,\mathrm{K}$ . Reading it: linear in $X_{\mathrm{CNO}}$ confirms the catalytic role, while the huge exponent $\nu$ concentrates all the burning at the centre. That exponent is no fitted constant; it comes from the thermally averaged rate through a Coulomb barrier, controlled by the Gamow factor.

r_{ij}=n_i n_j\langle\sigma v\rangle_{ij},\qquad \langle\sigma v\rangle\propto \exp\!\left[-3\left(\frac{b^2}{4k_BT}\right)^{1/3}\right],\;\; b=\frac{\pi\sqrt{2\mu}\,Z_1Z_2 e^2}{\hbar} \qquad (5)

where: $n_i,n_j$ — number densities of the reacting species; $\langle\sigma v\rangle$ — thermally averaged rate per pair; $b$ — Gamow energy factor, growing linearly with $Z_1Z_2$ ; $\mu$ — reduced mass. Reading it: because $b\propto Z_1Z_2=6$ – $8$ against $1$ for two protons, the tunnelling exponent is much larger, and its temperature derivative produces $\nu\approx18$ . In steady state the loop settles into a characteristic distribution: writing $\lambda_k$ for the proton-capture destruction rate of species $k$ , the equilibrium abundances are fixed by a single conserved flux around the loop.

\frac{n_i}{n_j}=\frac{\lambda_j}{\lambda_i},\qquad \lambda_k=n_p\langle\sigma v\rangle_{p,k} \qquad (6)

where: $n_i$ — equilibrium number density of catalyst $i$ ; $\lambda_k$ — its proton-capture destruction rate; $n_p$ — proton number density. Reading it: abundance is inversely proportional to destruction rate, so the slowest-reacting species hosts the most nuclei — applied to the slow $^{14}\mathrm{N}(p,\gamma)$ step it gives $\big(n(^{12}\mathrm{C})/n(^{14}\mathrm{N})\big)_{\mathrm{eq}}=\lambda_{^{14}\mathrm{N}}/\lambda_{^{12}\mathrm{C}}\ll 1$ , the origin of the nitrogen enhancement and low $^{12}\mathrm{C}/^{13}\mathrm{C}$ ratio in dredged-up material. Finally, the structural payoff: CNO-powered stars populate the upper main sequence and obey a steeper mass–luminosity relation than low-mass pp stars, which sets their lifetimes.

L\propto M^{3\text{–}4},\qquad t_{\mathrm{MS}}\propto\frac{M}{L}\propto M^{-2\text{–}3} \qquad (7)

where: $L$ — stellar luminosity; $M$ — stellar mass; $t_{\mathrm{MS}}$ — main-sequence lifetime. Reading it: a $10\,M_\odot$ star is thousands of times more luminous than the Sun and burns its fuel roughly a hundred times faster — a direct consequence of CNO-dominated cores.

Derivation ATemperature sensitivity of the CNO cycle

The goal is to derive, and evaluate by hand, the celebrated exponent $\nu\approx 18$ in $\varepsilon_{\mathrm{CNO}}\propto T^{\nu}$ purely from Gamow-peak physics. We start from the thermally averaged rate — the convolution of the Maxwell–Boltzmann tail with the quantum tunnelling probability — locate the Gamow peak by steepest descent, reduce the rate to a form governed by one dimensionless number $\tau$ , and take its logarithmic temperature derivative. The point is that the cycle's enormous sensitivity is no accident but a rigid consequence of the nuclear charges involved.

Step 1. Write the thermally averaged rate as $\langle\sigma v\rangle=(8/\pi\mu)^{1/2}(k_BT)^{-3/2}\int_0^\infty S(E)\,e^{-E/k_BT-b/\sqrt{E}}\,\mathrm{d}E$ , where $S(E)$ is the slowly varying astrophysical S-factor and the exponential carries the competition between the Boltzmann factor (few particles at high $E$ ) and the tunnelling factor (poor penetration at low $E$ ). Why: isolating the smooth $S(E)$ puts all the temperature dependence into the sharply peaked exponential.

Step 2. The two competing exponentials produce a sharp maximum, the Gamow peak, at $E_0=(b\,k_BT/2)^{2/3}$ ; steepest descent about $E_0$ gives a rate controlled by the single dimensionless Gamow exponent $\tau\equiv 3E_0/k_BT=3\,(b^2/4k_BT)^{1/3}$ . Why: the saddle-point value collapses all the barrier physics into $\tau$ , and since $b\propto Z_1Z_2\sqrt{\mu}$ , this is where the target charge enters.

Step 3. After the Gaussian integration the rate reduces to $\langle\sigma v\rangle\propto \tau^{2}e^{-\tau}$ . Why: this compact form differentiates cleanly — the $\tau^2$ prefactor is a small correction and $e^{-\tau}$ carries the steep dependence.

Step 4. Since $\tau\propto T^{-1/3}$ , $\mathrm{d}\tau/\mathrm{d}\ln T=-\tau/3$ , so $\nu=\dfrac{\mathrm{d}\ln\langle\sigma v\rangle}{\mathrm{d}\ln T}=\Big(\dfrac{2}{\tau}-1\Big)\!\Big(-\dfrac{\tau}{3}\Big)=\dfrac{\tau-2}{3}$ . Why: the index is set entirely by $\tau$ — a bigger barrier means a steeper temperature dependence, the mechanism behind the whole article.

Step 5. Evaluate for $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ at $T=2\times10^{7}\,\mathrm{K}$ : with $Z_1Z_2=7$ , $\mu\approx 0.93\,m_u$ , the Gamow energy is $b^2\approx 44.8\,\mathrm{MeV}$ and $k_BT\approx 1.72\,\mathrm{keV}$ , so $\tau\approx 56$ and $\nu=(56-2)/3\approx 18$ . Why: the same computation for proton–proton fusion ( $Z_1Z_2=1$ ) gives $\tau\approx 14$ , $\nu\approx 4$ — the origin of the vastly different sensitivities of the two burning modes. *Verified: reproduces Carroll & Ostlie's $\varepsilon_{\mathrm{CNO}}\propto T^{\sim 19.9}$ at $1.5\times10^{7}\,\mathrm{K}$ (recomputed here as $\tau\approx 62$ , $\nu\approx 19.9$ ) and $\sim 18$ at $2\times10^{7}\,\mathrm{K}$ ; consistent with Clayton Ch. 5.*

Go deeper · Derivation B · 1 minEquilibrium abundances and the nitrogen bottleneck

Derivation BEquilibrium abundances and the nitrogen bottleneck

The goal is to prove that in CN-cycle equilibrium the catalyst abundances arrange themselves inversely to their proton-capture rates, so the slowest reaction hosts the most nuclei and carbon is systematically converted into nitrogen. We write the number-density rate equations, impose steady state, extract the abundance ratios, and identify the bottleneck — the reasoning that turns a kinetics statement into the surface-abundance anomalies of evolved stars.

Step 1. For each catalyst $k$ write $\dot n_k=\lambda_{k-1}n_{k-1}-\lambda_k n_k$ , where $\lambda_k=n_p\langle\sigma v\rangle_{p,k}$ : species $k$ is created by proton capture on its predecessor and destroyed by proton capture on itself. Why: every gain for one species is a loss for the one before it, so the equations couple into a single closed cycle.

Step 2. Impose steady state, $\dot n_k=0$ , which forces $\lambda_k n_k=\lambda_{k-1}n_{k-1}=\dots=F$ , a single loop flux $F$ shared by every reaction. Why: in equilibrium the same number of nuclei must pass through each step per second, or material would pile up — the loop is a series of reservoirs draining at identical rate $F$ .

Step 3. Solve for the abundances: $n_k=F/\lambda_k$ , hence $n_i/n_j=\lambda_j/\lambda_i$ . Why: a species with a small destruction rate needs a large population to carry the common flux $F$ , so abundance and reaction speed are reciprocal.

Step 4. Identify the bottleneck: $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ has the smallest cross section, so $\lambda_{^{14}\mathrm{N}}$ is smallest and $n(^{14}\mathrm{N})$ largest; almost all catalytic $^{12}\mathrm{C}$ is driven into $^{14}\mathrm{N}$ , and $^{12}\mathrm{C}/^{13}\mathrm{C}$ falls from its solar value near 89 toward a few. Why: this converts one slow cross section into the observable nitrogen enhancement and low carbon-isotope ratio that betray CNO cycling deep inside a star. *Verified: predicts the observed CN-equilibrium $^{12}\mathrm{C}/^{13}\mathrm{C}\approx 3$ – $4$ and strong nitrogen enhancement seen in CNO-processed material after first dredge-up (Kippenhahn, Weigert & Weiss §18).*

Go deeper · Derivation C · 1 minWhy steep burning builds a convective core

Derivation CWhy steep burning builds a convective core

The goal is to connect the $T^{18}$ sensitivity to a structural feature that reshapes the upper main sequence: the convective core of a massive star. We recall the Schwarzschild criterion and the radiative gradient a star needs to carry its luminosity, and show that when energy generation is concentrated at the centre, the required gradient there exceeds what an adiabatic gas can sustain — so the core boils, bridging nuclear kinetics to stellar hydrodynamics.

Step 1. State the Schwarzschild criterion: a layer is convectively unstable when the gradient needed to carry the flux by radiation exceeds the adiabatic gradient, $\nabla_{\mathrm{rad}}>\nabla_{\mathrm{ad}}$ , with $\nabla_{\mathrm{ad}}\approx 0.4$ for an ideal gas. Why: this switch decides whether a region transports energy by photon diffusion or bulk convection.

Step 2. The radiative gradient needed to carry the local luminosity is $\nabla_{\mathrm{rad}}=\dfrac{3}{16\pi acG}\dfrac{\kappa\,L(r)\,P}{m(r)\,T^{4}}$ , with $L(r)$ and $m(r)$ the luminosity and mass enclosed within $r$ . Why: the more energy pushed through a shell, the steeper the gradient radiation demands — so $\nabla_{\mathrm{rad}}$ tracks $L(r)/m(r)$ .

Step 3. Because $\varepsilon_{\mathrm{CNO}}\propto T^{18}$ , nearly all the luminosity is generated in a tiny central sphere where $L(r)/m(r)$ is largest, so $\nabla_{\mathrm{rad}}$ peaks at the centre and exceeds $\nabla_{\mathrm{ad}}$ for $M\gtrsim 1.3\,M_\odot$ : a convective core forms, mixing fresh hydrogen inward. In a pp star ( $\varepsilon\propto T^{4}$ ) the generation is spread out and the centre stays radiative. Why: the exponent $\nu$ sets how concentrated the burning is — splitting the main sequence into convective-core upper stars and radiative-core lower stars, downstream entirely of the CNO exponent. *Verified: consistent with the standard result that main-sequence stars above $\sim 1.1$ – $1.3\,M_\odot$ develop convective cores (Kippenhahn, Weigert & Weiss §22–23; Carroll & Ostlie §10.4).*

Limiting cases

The same rate law reads very differently across the temperature ladder — from the Sun's core, where the cycle idles, to nova explosions, where it saturates.

Case	Result	Meaning
Low temperature ( $T\lesssim 1.5\times10^{7}\,\mathrm{K}$ )	$\varepsilon_{\mathrm{CNO}}\ll\varepsilon_{pp}$	In the Sun's core the CNO cycle makes only about 1% of the luminosity; the pp chain dominates.
Crossover ( $T\approx 1.7\times10^{7}\,\mathrm{K}$ )	$\varepsilon_{\mathrm{CNO}}\simeq\varepsilon_{pp}$	Roughly the central temperature of a $1.3\,M_\odot$ star — the mass above which CNO burning takes over.
High temperature ( $T\gtrsim 2\times10^{7}\,\mathrm{K}$ )	$\varepsilon_{\mathrm{CNO}}\propto T^{\sim 18}$	Upper-main-sequence stars: burning is so centrally concentrated that a convective core forms.
Hot CNO / beta-limited ( $T\gtrsim 10^{8}\,\mathrm{K}$ )	$\varepsilon\ \text{set by}\ \tau_{1/2}(^{13}\mathrm{N},{}^{15}\mathrm{O})$	In novae and X-ray bursts proton captures outrun the beta decays; the rate saturates at fixed decay lifetimes.

Worked examples

1 · Deriving ν ≈ 18 for the nitrogen bottleneck (fully worked)

Confirm the temperature exponent of the CNO cycle by computing the Gamow exponent $\tau$ for the rate-limiting reaction $^{14}\mathrm{N}(p,\gamma)^{15}\mathrm{O}$ at a core temperature $T=2\times10^{7}\,\mathrm{K}$ , then use $\nu=(\tau-2)/3$ .

Step 1. Assemble the Gamow energy $b^{2}=2\mu c^{2}(\pi\alpha Z_1Z_2)^{2}$ with $Z_1Z_2=7$ , reduced mass number $14/15=0.933$ so $\mu c^{2}=869\,\mathrm{MeV}$ , and $\pi\alpha Z_1Z_2=0.1605$ : $b^{2}=2(869)(0.1605)^{2}=\mathbf{44.8\ MeV}$ .

Step 2. Thermal energy: $k_BT=(8.617\times10^{-5}\,\mathrm{eV\,K^{-1}})(2\times10^{7}\,\mathrm{K})=\mathbf{1.72\ keV}$ .

Step 3. Gamow exponent: $\tau=3\,(b^{2}/4k_BT)^{1/3}=3\,(44.8\times10^{6}/(4\times1723))^{1/3}=3\,(6496)^{1/3}=\mathbf{56.0}$ .

Step 4. Power-law index: $\nu=(\tau-2)/3=(56.0-2)/3=\mathbf{18.0}$ — the famous $\varepsilon_{\mathrm{CNO}}\propto T^{18}$ scaling, recovered from nuclear charges and temperature alone with no free parameters.

Step 5. Contrast with proton–proton fusion at the Sun's core ( $Z_1Z_2=1$ , $\mu=0.5\,m_u$ ): the same recipe gives $b^{2}=0.49\,\mathrm{MeV}$ , $\tau\approx 14$ , $\nu\approx\mathbf{4}$ . The ratio of exponents, 18 versus 4, is the whole reason the two burning modes divide the main sequence between them.

2 · The neutrino tax on CNO energy (fully worked)

The net $Q=26.73\,\mathrm{MeV}$ per He-4 is the same for the CNO cycle and the pp chain, but the CNO cycle emits two beta-decay neutrinos with mean energies $\langle E_\nu(^{13}\mathrm{N})\rangle=0.71\,\mathrm{MeV}$ and $\langle E_\nu(^{15}\mathrm{O})\rangle=1.00\,\mathrm{MeV}$ . What fraction of the released energy escapes as neutrinos, and how much is deposited as heat?

Step 1. Total neutrino energy per cycle: $E_\nu=0.71+1.00=\mathbf{1.71\ MeV}$ ; fraction lost $f_\nu=E_\nu/Q=1.71/26.73=\mathbf{6.4\%}$ .

Step 2. Energy thermalized in the star: $Q_{\mathrm{eff}}=Q-E_\nu=26.73-1.71=\mathbf{25.0\ MeV}$ .

Step 3. The pp chain's softer neutrinos carry away only about 2%, so CNO is a slightly "leakier" reactor — for the same nuclear budget roughly 4% less energy holds the star up. CNO loses ≈ 6.4% to neutrinos versus ≈ 2% for the pp chain, depositing 25.0 MeV of heat per helium-4.

Physics of the portrait

The CNO cycle is where nuclear physics first explained why the stars shine — a discovery of the late 1930s, when proton-capture cross sections were finally good enough to test against the Sun and the hotter, brighter stars the pp chain alone could not power.

Origins · 1 minHans Bethe and how this was discovered

Hans Bethe

1906–2005 · nuclear astrophysics and stellar energy generation

Hans Bethe worked out the CNO cycle in 1938, one of the most consequential calculations of the century. Trained in Munich under Sommerfeld and driven from Germany in 1933, Bethe had become the master of nuclear reaction theory — his three-part "Bethe Bible" in Reviews of Modern Physics codified the young field. At a small 1938 Washington conference organized by Gamow and Teller, the astrophysicists posed the problem: what reactions could power a star? Within weeks Bethe had identified two solutions — the proton–proton chain for cooler stars like the Sun, and the catalytic carbon–nitrogen cycle for hotter, massive stars — crucially showing that the C, N, O nuclei act only as catalysts, so a trace abundance suffices. Carl Friedrich von Weizsäcker had independently glimpsed the cycle in Germany at nearly the same time, but Bethe's 1939 paper "Energy Production in Stars" gave the quantitative theory, matching the Sun's luminosity and predicting the steep temperature dependence. It won him the 1967 Nobel Prize "for his discoveries concerning the energy production in stars."

History

1920

— Eddington argues that subatomic energy, not gravitational contraction, powers the stars

1937

— von Weizsäcker begins developing catalytic nuclear cycles for stellar energy

1938

— Bethe formulates the CNO cycle and the competing proton–proton chain

1967

— Bethe awarded the Nobel Prize for the theory of stellar energy generation

2020

— Borexino reports the first direct detection of solar CNO neutrinos

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., §10.3 (nucleosynthesis and the CNO cycle) and §10.4 (stellar structure) · Clayton, Principles of Stellar Evolution and Nucleosynthesis, Ch. 5 (thermonuclear reaction rates and the Gamow peak) · Kippenhahn, Weigert & Weiss, Stellar Structure and Evolution, 2nd ed., §18 (nuclear energy production) and §22–23 (convection) · Iliadis, Nuclear Physics of Stars, Ch. 5 · Adelberger et al. 2011, Rev. Mod. Phys. 83, 195 (solar fusion cross sections).

Provenance. The net $Q=26.73\,\mathrm{MeV}$ and the six-reaction CN sequence are standard (Carroll & Ostlie §10.3; Clayton Ch. 5). The temperature exponent was recomputed here from first principles: $b^{2}=2\mu c^{2}(\pi\alpha Z_1Z_2)^{2}=44.8\,\mathrm{MeV}$ for $^{14}\mathrm{N}(p,\gamma)$ gives $\tau=56$ and $\nu=(\tau-2)/3=18.0$ at $2\times10^{7}\,\mathrm{K}$ (and $\nu\approx 19.9$ at $1.5\times10^{7}\,\mathrm{K}$ , matching Carroll & Ostlie). The $\sim 6\%$ neutrino loss (mean $\nu$ energies $0.71$ and $1.00\,\mathrm{MeV}$ ), the crossover at $\sim 1.7\times10^{7}\,\mathrm{K}$ / $\sim 1.3\,M_\odot$ , the CN-equilibrium $^{12}\mathrm{C}/^{13}\mathrm{C}\approx 3$ – $4$ , the $L\propto M^{3\text{–}4}$ slope, and Borexino 2020 are all standard. A verifier should recompute $\tau$ and $\nu$ and check the equilibrium derivation against Kippenhahn, Weigert & Weiss §18.

Prerequisites. The Proton–Proton Chain · Nuclear Reaction Rates and the Gamow Peak · The Main Sequence. Leads to. The Mass–Luminosity Relation · Helium Burning and the Triple-Alpha Process · Post-Main-Sequence Evolution and the Red Giant Branch · Advanced Burning Stages in Massive Stars.

Ask the tutor: Ask me to walk through the Gamow-exponent arithmetic for any reaction, or to trace how the nitrogen bottleneck turns into the surface abundances we see in red giants.

If C, N, and O are only catalysts, why do stars need any of them to run the cycle?

Because a catalyst, though conserved, must be present to provide the low-barrier pathway. The trick of the CNO cycle is that fusing a proton onto an existing carbon nucleus is easier at high temperature than the awkward, weak-interaction-limited first step of the pp chain. This is why the very first stars — Population III, formed from primordial gas with essentially zero metals — could not use the cycle and had to contract to much higher central temperatures, until they made a little carbon by the triple-alpha process to prime it.

Why is ¹⁴N(p,γ)¹⁵O the slowest step, and does it really control the whole cycle?

Yes — the slowest link sets the pace of the entire loop, just as the slowest booth sets a toll plaza's throughput. The reaction is slow because it is a non-resonant radiative capture with a small astrophysical S-factor, and nitrogen's charge makes its barrier high. In steady state the cycle's overall rate is essentially the rate of this one reaction, which is why measuring its cross section precisely (the job of experiments like LUNA) sets CNO energy generation, globular-cluster ages, and the predicted solar CNO-neutrino flux.

Why does the switch from pp to CNO happen over such a narrow mass range?

Because the two rates depend on temperature with wildly different steepness — near $1.5$ – $2\times10^{7}\,\mathrm{K}$ the pp rate grows as $T^{4}$ while CNO grows as $T^{18}$ . Raising the central temperature by just 33% (from $1.5$ to $2.0\times10^{7}\,\mathrm{K}$ ) multiplies the CNO-to-pp ratio by $1.333^{18}/1.333^{4}\approx 56$ . Since central temperature is a gentle function of stellar mass, that explosive sensitivity tips the balance from pp- to CNO-dominated over a mass interval of only tens of percent, centred near $1.3\,M_\odot$ .

The net reaction and Q-value are identical to the pp chain — so why does it matter which runs?

The net transmutation and the $26.73\,\mathrm{MeV}$ are the same, but three things differ. The neutrino losses are larger for CNO ( $\sim 6\%$ versus $\sim 2\%$ ), so slightly less is deposited as heat. Far more important, the extreme temperature sensitivity concentrates burning into a pinpoint at the centre, driving a convective core that mixes fuel and alters evolution. And the cycle rearranges the star's C, N, O into equilibrium ratios that later surface as abundance anomalies. Same energy, very different star.

What is the "hot CNO cycle," and where does it operate?

Above about $10^{8}\,\mathrm{K}$ — reached in classical novae on white dwarfs and X-ray bursts on neutron stars — proton captures become so fast that they outrun the beta decays of $^{13}\mathrm{N}$ and $^{15}\mathrm{O}$ . The cycle stops being temperature-limited and becomes "beta-limited": its rate is fixed by those half-lives rather than by how hot the gas is. Energy generation saturates, changing the explosion physics, and the frozen-in radioactive nuclei enrich the ejecta in $^{13}\mathrm{C}$ , $^{15}\mathrm{N}$ , and $^{17}\mathrm{O}$ — signatures seen in meteoritic stardust grains.

How do we know the CNO cycle happens inside stars we cannot see into?

Three independent lines of evidence. Direct: Borexino detected the specific neutrinos from the $^{13}\mathrm{N}$ and $^{15}\mathrm{O}$ beta decays in the Sun in 2020, a signal that can only come from the cycle. Chemical: the surfaces of evolved intermediate-mass stars show exactly the nitrogen enhancement and low $^{12}\mathrm{C}/^{13}\mathrm{C}$ ratios that CN equilibrium predicts, revealed when convection dredges core-processed material upward. Structural: the upper main sequence follows the steep mass–luminosity relation and convective-core signatures that only a $T^{18}$ source can produce.