Energy Generation Equation

A star in a steady state is a machine that manufactures energy in a tiny, blazing core and delivers it, kilowatt by kilowatt, to a vast surface that radiates it into space. The energy generation equation is the accountant's ledger for that delivery: it says that as you move outward from radius $r$ to $r+\mathrm{d}r$ , the interior luminosity $L_r$ — the net power crossing the sphere of radius $r$ — increases by exactly the energy the intervening shell produces per second, no more and no less. Because the shell carries mass $\mathrm{d}m = 4\pi r^2\rho\,\mathrm{d}r$ and each kilogram liberates power $\varepsilon$ , the bookkeeping closes as $\mathrm{d}L_r = \varepsilon\,\mathrm{d}m$ . This is nothing but energy conservation written in spherical symmetry. It is the third of the four coupled equations of stellar structure — alongside the Mass Continuity Equation, Hydrostatic Equilibrium in Stars, and the temperature-gradient equation — but it is the one that ushers physics through the door. Mass continuity and hydrostatic balance are purely mechanical; the energy equation drags in the microphysics of nuclear reactions, neutrino losses, and thermodynamics through the single symbol $\varepsilon$ , coupling the run of luminosity to the composition, temperature, and density of the plasma. The subtlety lives in $\varepsilon$ . In a main-sequence star it is dominated by thermonuclear burning, whose ferocious temperature sensitivity — from quantum tunnelling through the Coulomb barrier — confines almost all energy release to the central few percent of the radius. But $\varepsilon$ is a net specific power: it also carries a gravothermal term, positive in contraction and negative in storage, and a neutrino term that drains energy the moment it is made. Integrated over the star, $L = \int_0^M \varepsilon\,\mathrm{d}m$ fixes the surface luminosity and, with the other structure equations, the whole theory of stellar models.

Key takeaways

The energy generation equation $\mathrm{d}L_r/\mathrm{d}r = 4\pi r^2\rho\,\varepsilon$ says the interior luminosity grows outward at the rate each spherical shell liberates energy — pure energy conservation in spherical symmetry.
In mass coordinates it collapses to the compact Lagrangian form $\mathrm{d}L_r/\mathrm{d}m = \varepsilon$ , the version stellar-evolution codes actually integrate.
The specific rate splits as $\varepsilon = \varepsilon_{\mathrm{nuc}} - \varepsilon_\nu + \varepsilon_{\mathrm{grav}}$ : nuclear burning, minus freely-escaping neutrinos, plus the gravothermal source $-T\,\mathrm{d}s/\mathrm{d}t$ .
The steep law $\varepsilon_{\mathrm{nuc}}\propto\rho\,T^\nu$ ( $\nu\approx4$ for the p–p chain, $\nu\approx16$ – $18$ for CNO) crushes energy generation into a compact core: in the Sun half of $L$ is made inside $0.1\,R_\odot$ .
Integrating gives the surface luminosity $L=\int_0^M\varepsilon\,\mathrm{d}m$ ; for the Sun, $L_\odot=3.828\times10^{26}$ W, sustained by fusing about $6\times10^{11}$ kg of hydrogen every second — while the gravothermal term powers contraction and cooling and can even make $L_r$ decrease outward when neutrino losses win.
Key idea: energy conservation in spherical symmetry — the luminosity crossing radius $r$ rises outward at exactly the rate each shell liberates energy, $\mathrm{d}L_r/\mathrm{d}r = 4\pi r^2\rho\,\varepsilon$ , where $\varepsilon$ is the net specific power from nuclear reactions minus neutrino losses plus gravitational release.

The core insight

Energy conservation in spherical symmetry: the luminosity crossing radius $r$ increases outward at exactly the rate at which each shell liberates energy, $\mathrm{d}L_r/\mathrm{d}r = 4\pi r^2\rho\,\varepsilon$ , where $\varepsilon$ is the net specific power from nuclear reactions minus neutrino losses plus gravitational release.

Where this lives today

The energy generation equation is one of four lines integrated millions of times inside every modern stellar-evolution code — MESA, the Geneva and Padova grids — that predict the tracks stars trace across the Hertzsprung–Russell diagram. Its source term $\varepsilon$ is now assembled from tabulated, experimentally measured reaction rates (the REACLIB and NACRE compilations) rather than a single power law. The equation's deepest modern test is the Sun: helioseismology maps the internal sound speed to a fraction of a percent and confirms the compact burning core the equation predicts, while the Borexino experiment has directly detected neutrinos from both the p–p chain and the CNO cycle — measuring the very $\varepsilon_\nu$ term the equation subtracts. At the frontier, the neutrino-dominated regime where $\varepsilon_\nu$ drives $\mathrm{d}L_r/\mathrm{d}r<0$ governs a massive star's final days and sets the stage for core-collapse supernovae, where the coupling to three-dimensional convection remains genuinely open.

The physical picture

Picture the star as a stack of nested spherical shells and energy as water poured from below. Each shell adds its own trickle $\varepsilon\,\mathrm{d}m$ to a stream of luminosity flowing steadily upward. The flow crossing any radius is the sum of everything generated beneath it, so $L_r$ can only rise — or stay flat where nothing is generated; it never appears from nowhere. The first figure freezes one shell and reads its energy balance off the geometry.

Figure

A star as nested shells: a shell of mass dm = 4πr²ρ dr adds εdm to the outward flow, so the interior luminosity climbs from the core. In the Sun the burning core sits at T_c ≈ 1.5×10⁷ K and reaches only to about 0.2 R⊙

Because $\varepsilon_{\mathrm{nuc}}$ climbs as a very high power of temperature, energy release is savagely concentrated where $T$ is greatest — dead centre. Outside the burning core $\varepsilon$ falls essentially to zero, $L_r$ stops growing, and the envelope becomes a passive conduit carrying the fixed luminosity outward. The equation is the local face of the global truth $L=\int\varepsilon\,\mathrm{d}m$ : if the fusion the core supplies does not match the luminosity the surface demands, the star is out of thermal equilibrium and must expand or contract until $\varepsilon_{\mathrm{grav}}$ balances the books. The second figure shows this concentration for a real star.

Figure

Interior luminosity of a standard solar model: L_r rises steeply through the burning core and plateaus at L⊙ = 3.828×10²⁶ W. Half of L is generated inside 0.1 R⊙ (about 0.1% of the volume) and 99% within 0.25 R⊙; beyond that the envelope only transports it

The core equations

The equation is a local statement of energy conservation. Step outward by $\mathrm{d}r$ and $L_r$ grows by exactly what the intervening shell makes per second. The shell's mass is $4\pi r^2\rho\,\mathrm{d}r$ and each kilogram releases power $\varepsilon$ , so dividing by $\mathrm{d}r$ gives the Eulerian (radius) form.

\frac{\mathrm{d}L_r}{\mathrm{d}r} = 4\pi r^2\rho\,\varepsilon \qquad (1)

where: $L_r$ — interior luminosity, the net rate of energy flow across the sphere of radius $r$ (W); $r$ — radial coordinate from the centre (m); $\rho$ — local mass density (kg m⁻³); $\varepsilon$ — total specific energy generation rate, energy per unit mass per unit time (W kg⁻¹). Where $\varepsilon=0$ the luminosity is flat. Stellar-evolution codes prefer the enclosed mass $m$ as variable, since mass is conserved while radius is not. Dividing equation (1) by $\mathrm{d}m/\mathrm{d}r = 4\pi r^2\rho$ cancels the geometric factor and leaves the strikingly simple Lagrangian form.

\frac{\mathrm{d}L_r}{\mathrm{d}m} = \varepsilon \qquad (2)

where: $m$ — mass enclosed within radius $r$ , the natural independent variable for stellar evolution (kg); $\varepsilon$ — total specific energy generation rate (W kg⁻¹). The luminosity added per unit mass is just the energy each unit mass makes per second. The symbol $\varepsilon$ is not a single process but a net budget: nuclear reactions feed it, thermal neutrinos drain it (they escape at once, carrying their energy away), and a gravothermal term adds or subtracts as the gas releases or stores heat. Only their sum enters the structure equation.

\varepsilon = \varepsilon_{\mathrm{nuc}} - \varepsilon_\nu + \varepsilon_{\mathrm{grav}} \qquad (3)

where: $\varepsilon_{\mathrm{nuc}}$ — specific power from thermonuclear reactions (W kg⁻¹); $\varepsilon_\nu$ — specific power lost to freely escaping thermal neutrinos (W kg⁻¹); $\varepsilon_{\mathrm{grav}}$ — specific gravothermal power (W kg⁻¹), positive when the gas gives up heat. Neutrinos carry a minus sign because, unlike photons, they leave the accounting immediately rather than joining the outgoing luminosity. The gravothermal term is where thermodynamics enters — the rate at which each gram surrenders heat, which the first and second laws let us write as an entropy rate, or as internal-energy change plus compressional work.

\varepsilon_{\mathrm{grav}} = -T\frac{\mathrm{d}s}{\mathrm{d}t} = -\frac{\mathrm{d}u}{\mathrm{d}t} + \frac{P}{\rho^2}\frac{\mathrm{d}\rho}{\mathrm{d}t} \qquad (4)

where: $s$ — specific entropy (J K⁻¹ kg⁻¹); $u$ — specific internal energy (J kg⁻¹); $P$ — pressure (Pa); $T$ — temperature (K). When the gas contracts, its entropy falls ( $\mathrm{d}s/\mathrm{d}t<0$ ) and $\varepsilon_{\mathrm{grav}}>0$ : released energy acts as a genuine luminosity source. The nuclear rate follows from tunnelling through the Coulomb barrier averaged over a thermal distribution, and near any operating temperature is well approximated by a power law. Its temperature exponent is huge — the single most important fact about $\varepsilon_{\mathrm{nuc}}$ .

\varepsilon_{\mathrm{nuc}} = \varepsilon_0\,\rho\,X_a X_b\,T^\nu \qquad (5)

where: $\varepsilon_0$ — constant absorbing the cross-section and reaction physics; $X_a,X_b$ — mass fractions of the two reacting nuclei; $\nu$ — effective temperature exponent, $\nu\approx4$ for the proton–proton chain near $1.5\times10^7$ K and $\nu\approx16$ – $18$ for CNO. The linear $\rho$ reflects a two-body reaction; the enormous $T^\nu$ confines burning to the core. Integrating the Lagrangian form (2) from centre ( $L_0=0$ ) to surface gives the total power the star pours into space — its luminosity.

L = \int_0^M \varepsilon\,\mathrm{d}m, \qquad L_\odot = 3.828\times10^{26}\ \mathrm{W} \qquad (6)

where: $L$ — surface luminosity (W); $M$ — total stellar mass (kg); $L_\odot$ — the IAU nominal solar luminosity. Each fusion of four protons into one ${}^4\mathrm{He}$ releases $Q_{\mathrm{pp}}=26.73$ MeV, of which about $0.6$ MeV escapes as neutrinos; the rest becomes photon luminosity.

Derivation AEnergy generation from spherical energy conservation

The goal is to build equation (1) from first principles by applying energy conservation to a single static shell, converting a specific rate per kilogram into a luminosity gradient per metre. We work in spherical symmetry so every quantity depends only on $r$ , assume a steady state with no local energy storage, and take the shell thin enough that $\rho$ and $\varepsilon$ are constant across it — which makes the mandatory density factor transparent.

Step 1. Isolate the shell between $r$ and $r+\mathrm{d}r$ ; its volume is $\mathrm{d}V=4\pi r^2\,\mathrm{d}r$ and, by mass continuity, its mass is $\mathrm{d}m=\rho\,\mathrm{d}V=4\pi r^2\rho\,\mathrm{d}r$ . Why: energy is generated per unit mass, so the shell's mass — not its volume — multiplies $\varepsilon$ ; the $r^2$ factor encodes that outer shells hold far more mass.

Step 2. Write the steady-state balance: power out of the top equals power in at the bottom plus power generated inside, so $L_r(r+\mathrm{d}r)=L_r(r)+\varepsilon\,\mathrm{d}m$ . Why: in a steady state nothing accumulates, so what the shell makes shows up as extra luminosity leaving its outer face — energy conservation with no storage term.

Step 3. Substitute the shell mass to get $L_r(r+\mathrm{d}r)-L_r(r)=4\pi r^2\rho\,\varepsilon\,\mathrm{d}r$ , Taylor-expand the left side and divide by $\mathrm{d}r$ . Why: the difference over an infinitesimal step is the derivative times $\mathrm{d}r$ , so dividing isolates the gradient.

Step 4. Read off $\mathrm{d}L_r/\mathrm{d}r=4\pi r^2\rho\,\varepsilon$ , and change variables with $\mathrm{d}m=4\pi r^2\rho\,\mathrm{d}r$ to reach the Lagrangian twin $\mathrm{d}L_r/\mathrm{d}m=\varepsilon$ . Why: same physics, different clothing; the mass form is cleaner because the geometric factor cancels, so it is what evolution codes integrate. *Why the units work: the right-hand side carries (kg m⁻³)(m²)(W kg⁻¹) = W m⁻¹, confirming that $\mathrm{d}L_r/\mathrm{d}r$ is indeed a luminosity gradient.*

Go deeper · Derivation B · 1 minThe gravothermal term and the first law of thermodynamics

Derivation BThe gravothermal term and the first law of thermodynamics

The goal is to relax the steady-state assumption and recover the full time-dependent energy equation, in which a shell may store internal energy or do compressional work rather than pass all its heat straight through. This is what links stellar structure to thermodynamics: the extra term $\varepsilon_{\mathrm{grav}}=-T\,\mathrm{d}s/\mathrm{d}t$ powers a star before fusion ignites and drains one as it cools, and vanishes precisely in exact thermal equilibrium. We proceed from the first law per unit mass.

Step 1. Write the first law for one gram, $\mathrm{d}q=\mathrm{d}u+P\,\mathrm{d}(1/\rho)$ , where $q$ is heat added, $u$ the specific internal energy, and $1/\rho$ the specific volume. Why: the compressional term $P\,\mathrm{d}(1/\rho)$ is the work done under expansion or squeezing — the channel by which contraction converts gravitational energy into heat.

Step 2. Balance the shell's heat in mass coordinates: nuclear input minus neutrino loss minus luminosity divergence equals the heat-gain rate, $(\varepsilon_{\mathrm{nuc}}-\varepsilon_\nu)-\mathrm{d}L_r/\mathrm{d}m=\mathrm{d}q/\mathrm{d}t$ . Why: $-\mathrm{d}L_r/\mathrm{d}m$ is the net heat delivered by radiation and convection, so the budget is what the shell makes minus what it passes on minus what neutrinos steal.

Step 3. Substitute $\mathrm{d}q/\mathrm{d}t=T\,\mathrm{d}s/\mathrm{d}t=\mathrm{d}u/\mathrm{d}t-(P/\rho^2)\,\mathrm{d}\rho/\mathrm{d}t$ , using $\mathrm{d}(1/\rho)=-\rho^{-2}\mathrm{d}\rho$ . Why: as an entropy rate the sign is obvious — losing entropy ( $\mathrm{d}s/\mathrm{d}t<0$ ) means dumping heat, which reappears as luminosity.

Step 4. Define $\varepsilon_{\mathrm{grav}}\equiv-T\,\mathrm{d}s/\mathrm{d}t$ and rearrange to $\mathrm{d}L_r/\mathrm{d}m=\varepsilon_{\mathrm{nuc}}-\varepsilon_\nu-\mathrm{d}u/\mathrm{d}t+(P/\rho^2)\,\mathrm{d}\rho/\mathrm{d}t$ , matching equation (3). Why: the gravothermal source is a genuine contribution to $\varepsilon$ that can rival nuclear burning, and it vanishes when $\mathrm{d}s/\mathrm{d}t=0$ — exact thermal equilibrium. *Why the sign is right: $\varepsilon_{\mathrm{grav}}\to0$ in exact thermal equilibrium (where $\mathrm{d}s/\mathrm{d}t=0$ ) and $\varepsilon_{\mathrm{grav}}>0$ during contraction (where entropy falls), consistent with the physical picture.*

Go deeper · Derivation C · 1 minWhy nuclear burning is confined to a compact core

Derivation CWhy nuclear burning is confined to a compact core

The goal is to turn the steep power-law rate of equation (5) into a quantitative statement about where in the star energy is actually made, and so explain why $L_r$ rises almost vertically near the centre and then flatlines. The engine is the exponent $\nu$ : because $\varepsilon_{\mathrm{nuc}}\propto T^\nu$ with $\nu$ as large as $16$ for CNO, even a modest outward drop in temperature collapses the burning rate by orders of magnitude. We take density roughly constant near the centre and let temperature decline outward.

Step 1. From equation (5), the ratio of local to central rate is $\varepsilon_{\mathrm{nuc}}(r)/\varepsilon_{\mathrm{nuc}}(0)\approx[T(r)/T_c]^{\nu}$ once the slowly varying density and composition factors are absorbed. Why: $\rho$ and the mass fractions change gently compared with the violent $T^\nu$ term, so temperature alone drives the profile.

Step 2. Evaluate where temperature has fallen to $0.8\,T_c$ : for CNO burning $(0.8)^{16}\approx0.028$ , so barely $3\%$ of the central rate survives. Why: a mere $20\%$ temperature drop kills $97\%$ of the generation, so the burning "core" is simply where $T$ is within a whisker of $T_c$ .

Step 3. Contrast the p–p chain, $(0.8)^{4}\approx0.41$ : far gentler, so p–p burning spreads over a larger central region. Why: the smaller exponent lets low-mass stars burn over a broad core, whereas the CNO exponent packs massive-star burning so tightly that the flux drives convection.

Step 4. Feed this back into equation (1): $\mathrm{d}L_r/\mathrm{d}r$ is large only in the hot centre, so $L_r$ climbs steeply there and saturates at $L$ once $\varepsilon_{\mathrm{nuc}}\to0$ . Why: this is the profile in the second figure — the outer $\sim90\%$ of a star by radius adds nothing to its luminosity, merely transporting it. *Numerically: $(0.8)^{16}=0.0281$ and $(0.8)^4=0.410$ , confirming the sharp pp-versus-CNO contrast and the convective-core outcome for massive stars.*

Limiting cases

Which of the three terms in $\varepsilon$ dominates depends entirely on the evolutionary phase, and the character of the equation changes with it.

Case	Result	Meaning
Static nuclear burning	$\varepsilon \approx \varepsilon_{\mathrm{nuc}}$	Main-sequence star in thermal equilibrium: gravothermal term negligible, $L=\int\varepsilon_{\mathrm{nuc}}\,\mathrm{d}m$ is fixed purely by fusion.
Gravitational contraction	$\varepsilon \approx \varepsilon_{\mathrm{grav}} = -T\,\mathrm{d}s/\mathrm{d}t > 0$	Pre-main-sequence (Kelvin–Helmholtz) phase or white-dwarf cooling: released gravitational and thermal energy supplies the luminosity.
Neutrino-dominated cooling	$\mathrm{d}L_r/\mathrm{d}r < 0$	Advanced burning in massive stars: thermal neutrino losses exceed local nuclear input, so the net budget is negative and $L_r$ falls outward.
Envelope (no generation)	$L_r = \text{const} = L$	Above the core $\varepsilon\to0$ ; the equation reduces to $\mathrm{d}L_r/\mathrm{d}r=0$ and the luminosity is carried unchanged to the surface.

Common trap“ε is just the nuclear burning rate.”

It is not — $\varepsilon$ is the net specific power, $\varepsilon = \varepsilon_{\mathrm{nuc}} - \varepsilon_\nu + \varepsilon_{\mathrm{grav}}$ , and treating it as purely nuclear gives the wrong luminosity in every phase except static main-sequence burning. The gravothermal term $\varepsilon_{\mathrm{grav}}=-T\,\mathrm{d}s/\mathrm{d}t$ can be positive (contraction, cooling) or negative (heating, expansion) and dominates entirely before ignition and after burning ceases — it powered the Sun for its first $\sim30$ million years. The neutrino term is subtracted because thermal neutrinos leave instantly; in carbon burning and beyond they can make $\varepsilon<0$ so that $L_r$ decreases outward. A subtler trap: $L_r$ is the net flux, so in a convective zone it includes the convective enthalpy flux, not just radiation.

Worked examples

1 · The Sun's energy budget — how little, and how much

Given $L_\odot=3.828\times10^{26}$ W and $M_\odot=1.989\times10^{30}$ kg, find the Sun's mean specific generation rate, and — using $Q_{\mathrm{pp}}=26.73$ MeV per helium nucleus — the hydrogen mass fused and the mass converted to energy each second.

Step 1. Mean specific rate, $\bar\varepsilon=L_\odot/M_\odot=3.828\times10^{26}/1.989\times10^{30}=1.92\times10^{-4}$ W kg⁻¹. That is about $0.19$ mW per kilogram — a thousand times feebler than a human body's $\sim1$ W kg⁻¹ metabolism. The Sun shines only because it is enormous.

Step 2. Reactions per second: each $4\,{}^1\mathrm{H}\to{}^4\mathrm{He}$ liberates $Q_{\mathrm{pp}}=26.73\ \mathrm{MeV}=4.28\times10^{-12}$ J, so the Sun completes $L_\odot/Q_{\mathrm{pp}}=8.9\times10^{37}$ reactions each second, each consuming four protons — $3.6\times10^{38}$ protons per second.

Step 3. Hydrogen mass fused: $3.6\times10^{38}\times1.673\times10^{-27}=\mathbf{6.0\times10^{11}\ kg\,s^{-1}}$ — roughly 600 million tonnes every second. Of this, only the mass defect becomes energy: $L_\odot/c^2=3.828\times10^{26}/(2.998\times10^8)^2=\mathbf{4.3\times10^{9}\ kg\,s^{-1}}$ , about 4.3 million tonnes per second turned into sunlight. The Sun fuses 600 Mt s⁻¹ of hydrogen, converting 4.3 Mt s⁻¹ into radiant energy, yet averages only 0.19 mW kg⁻¹.

2 · How far out does burning reach?

A CNO core has $\nu\approx16$ and a p–p core has $\nu\approx4$ . At the radius where the temperature has dropped to $80\%$ of its central value, what fraction of the central generation rate survives in each?

Step 1. Use $\varepsilon_{\mathrm{nuc}}(r)/\varepsilon_{\mathrm{nuc}}(0)\approx[T(r)/T_c]^{\nu}$ with $T/T_c=0.8$ . For CNO: $(0.8)^{16}=\mathbf{0.028}$ , so only $\sim3\%$ of the central rate remains — burning is all but extinguished by a $20\%$ temperature drop.

Step 2. For the p–p chain: $(0.8)^{4}=\mathbf{0.41}$ , so $\sim41\%$ survives — a far gentler falloff, generating energy over a much broader central region than CNO does.

Step 3. Consequence: the CNO flux is crushed into so small a volume that the radiative gradient needed to carry it exceeds the adiabatic value, so massive stars develop convective cores; the Sun's gentler p–p burning leaves its core radiative. A 20% temperature drop cuts CNO generation to 3% but p–p to only 41% — which is why massive stars have convective cores and the Sun does not.

3 · Why gravity alone cannot power the Sun

Estimate the Kelvin–Helmholtz (gravothermal) timescale $t_{\mathrm{KH}}\sim GM^2/(RL)$ for the Sun and compare it with the nuclear timescale, taking $R_\odot=6.96\times10^8$ m and assuming $\sim10\%$ of the hydrogen burns at $0.7\%$ efficiency.

Step 1. Gravothermal store: $GM_\odot^2/R_\odot=(6.674\times10^{-11})(1.989\times10^{30})^2/(6.96\times10^8)=3.8\times10^{41}$ J. Dividing by $L_\odot$ gives $t_{\mathrm{KH}}=9.9\times10^{14}$ s $\approx\mathbf{3\times10^{7}\ yr}$ : running on contraction alone ( $\varepsilon\approx\varepsilon_{\mathrm{grav}}$ ), the Sun would exhaust its gravitational energy in only $\sim30$ million years.

Step 2. Nuclear store: $0.1\times0.007\times M_\odot c^2=1.25\times10^{44}$ J, giving $t_{\mathrm{nuc}}=1.25\times10^{44}/3.828\times10^{26}=3.3\times10^{17}$ s $\approx\mathbf{1\times10^{10}\ yr}$ .

Step 3. The ratio $t_{\mathrm{nuc}}/t_{\mathrm{KH}}\approx300$ is the historical clincher: geology demanded a Sun billions of years old, but gravity offered only tens of millions. Only the nuclear term in $\varepsilon$ — a $\sim10^{10}$ yr reservoir versus gravity's $\sim3\times10^7$ yr — can explain the Sun's age, exactly the argument Eddington pressed in the 1920s.

Physics of the portrait

The energy generation equation is the point where the mechanical skeleton of a star acquires a beating heart. Its author, more than anyone, was the man who insisted stars must run on subatomic energy decades before the reactions were known — and who wrote the equations of stellar structure into their modern form.

Origins · 1 minArthur Eddington and how this was discovered

Arthur Eddington

1882–1944 · stellar structure and the source of stellar energy

Arthur Stanley Eddington was the architect of stellar-interior theory. In his 1920 British Association address and his 1926 masterwork The Internal Constitution of the Stars, he set out the coupled equations — hydrostatic support, mass distribution, energy transport, and energy generation — that still frame every stellar model. His boldest stroke concerned the source term $\varepsilon$ . Kelvin and Helmholtz had argued gravitational contraction powered the Sun, but that gave only tens of millions of years, contradicting the geological age of the Earth. Eddington countered that the Sun must draw on "subatomic" energy — hydrogen fusing to helium, releasing the mass defect Aston had just measured — famously telling doubters to "go and find a hotter place." He could not yet write $\varepsilon_{\mathrm{nuc}}$ ; that awaited quantum tunnelling (Gamow, 1928) and the reaction networks of Bethe, Critchfield, and von Weizsäcker at the close of the 1930s, which supplied the proton–proton chain and the CNO cycle. But Eddington had already put the term in the equation and insisted the universe honour it. The energy generation equation is his conservation law with the microphysics filled in.

History

1920

— Eddington argues the Sun is powered by subatomic energy, not gravitational contraction

1926

— Eddington's The Internal Constitution of the Stars codifies the stellar-structure equations

1928

— Gamow explains barrier penetration by quantum tunnelling, making fusion at stellar temperatures possible

1938

— Bethe and Critchfield compute the proton–proton chain rate

1939

— Bethe publishes the CNO cycle in Energy Production in Stars, awarded the Nobel Prize in 1967

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 10 (The Interiors of Stars) · Kippenhahn, Weigert & Weiss, Stellar Structure and Evolution, 2nd ed., Ch. 4 · Prialnik, An Introduction to the Theory of Stellar Structure and Evolution, Ch. 5 · Hansen, Kawaler & Trimble, Stellar Interiors, 2nd ed., Ch. 1 · Clayton, Principles of Stellar Evolution and Nucleosynthesis, Ch. 6. Key anchor values used throughout: the shell energy balance gives $\mathrm{d}L_r/\mathrm{d}r=4\pi r^2\rho\,\varepsilon$ with units W m⁻¹; the first-law manipulation gives $\varepsilon_{\mathrm{grav}}=-T\,\mathrm{d}s/\mathrm{d}t$ , which vanishes in thermal equilibrium; $L_\odot=3.828\times10^{26}$ W (IAU 2015 nominal), $Q_{\mathrm{pp}}=26.73$ MeV, p–p $\nu\approx4$ , CNO $\nu\approx16$ – $18$ ; and the worked figures $\bar\varepsilon=1.9\times10^{-4}$ W kg⁻¹, $6\times10^{11}$ kg s⁻¹ hydrogen fused, $4.3\times10^{9}$ kg s⁻¹ converted to energy, $t_{\mathrm{KH}}\approx3\times10^7$ yr, $t_{\mathrm{nuc}}\approx10^{10}$ yr, and $(0.8)^{16}\approx0.028$ .

Prerequisites. Mass Continuity Equation · Hydrostatic Equilibrium in Stars · Equation of State for Stellar Matter · Virial Theorem for Stars. Leads to. Equations of Stellar Structure · Radiative Temperature Gradient · Convective Energy Transport.

Ask the tutor: ask me to walk through how the four stellar-structure equations close into a single boundary-value problem, or to show exactly where the gravothermal term takes over as a star leaves the main sequence.

Why is the density ρ in the equation but not in the Lagrangian form dL_r/dm = ε?

The two forms are the same physics with different independent variables. In the radius form you ask "how much does $L_r$ change per metre?", and the answer must count how much mass sits in that metre-thick shell — which is $4\pi r^2\rho$ , hence the density. In the mass form you ask "how much per kilogram?", and the answer is simply $\varepsilon$ . The factor $4\pi r^2\rho$ is exactly $\mathrm{d}m/\mathrm{d}r$ , so it cancels when you switch variables. Codes prefer the mass form because mass labels a fixed parcel of gas even as the star expands and contracts.

What actually is ε, and why can it be negative?

$\varepsilon$ is the net specific power — the total energy each kilogram of gas adds to (or removes from) the outgoing luminosity per second. It has three parts: nuclear burning $\varepsilon_{\mathrm{nuc}}$ , always positive; the gravothermal term $\varepsilon_{\mathrm{grav}}=-T\,\mathrm{d}s/\mathrm{d}t$ , positive when the gas releases heat and negative when it stores heat; and the neutrino loss $\varepsilon_\nu$ , subtracted because those neutrinos escape immediately. The sum can be negative — in the late burning stages of a massive star, thermal neutrino emission can be so intense that a shell loses more than fusion supplies, and then $L_r$ must decrease outward.

Why is nearly all of a star's energy made in such a tiny central region?

Because the nuclear rate depends on temperature as an enormous power, $\varepsilon_{\mathrm{nuc}}\propto T^\nu$ with $\nu$ between about 4 and 18. Temperature falls outward from the centre, and raising a declining number to a large power makes the result plummet: for the CNO cycle ( $\nu\approx16$ ), a mere 20% temperature drop leaves only 3% of the central rate. So the "burning core" is not a structural wall — it is simply the small central volume where the temperature is within a whisker of its peak. Everything outside merely transports the luminosity the core has already made, which is why $L_r$ climbs steeply near the centre and then plateaus.

How does this equation connect to the other equations of stellar structure?

It is the third of four coupled equations. Mass continuity gives $\mathrm{d}m/\mathrm{d}r=4\pi r^2\rho$ ; hydrostatic equilibrium gives the pressure gradient; the energy generation equation supplies $L_r$ ; and the temperature-gradient equation says how steep a temperature drop is needed to carry that $L_r$ outward. They close into a solvable boundary-value problem once you add an equation of state, opacities, and reaction rates. The energy equation is the crucial coupling because $\varepsilon$ depends on the temperature and density the others set — so the microphysics of fusion feeds back into the whole structure of the star.

If the gravothermal term exists, why did Kelvin's contraction theory fail?

Kelvin and Helmholtz were not wrong that gravitational contraction releases energy — the term $\varepsilon_{\mathrm{grav}}$ is real and does power stars during some phases. They were wrong about the magnitude of the reservoir. The Sun's total gravitational energy, spread over its present luminosity, lasts only about 30 million years (the Kelvin–Helmholtz timescale), but geology and radioactive dating showed the Earth to be billions of years old — a factor of a few hundred too long for gravity alone. That discrepancy is exactly why Eddington argued a far larger reservoir, subatomic energy, had to feed $\varepsilon$ . Gravitational contraction powers the approach to the main sequence; nuclear burning powers the long main-sequence life.

Does the equation change for a star that isn't in a steady state?

Yes, and gracefully. The steady-state version $\mathrm{d}L_r/\mathrm{d}m=\varepsilon_{\mathrm{nuc}}-\varepsilon_\nu$ assumes each shell passes on exactly what it makes. When a star contracts, expands, or cools, a shell also stores or releases heat, and the full equation adds the gravothermal term: $\mathrm{d}L_r/\mathrm{d}m=\varepsilon_{\mathrm{nuc}}-\varepsilon_\nu-\mathrm{d}u/\mathrm{d}t+(P/\rho^2)\,\mathrm{d}\rho/\mathrm{d}t$ . This general form governs pre-main-sequence contraction, thermal pulses, and white-dwarf and neutron-star cooling, where fusion may be absent and $\varepsilon_{\mathrm{grav}}$ carries the whole luminosity. In exact thermal equilibrium the extra term vanishes and you recover the static equation.