Mass-Luminosity Relation

The mass–luminosity relation (MLR) is the single most productive regularity in stellar astrophysics: for a star fusing hydrogen in its core, the total power it radiates is fixed, to first approximation, by its mass alone. Plot the luminosities of stars with dynamically measured masses against those masses and the points fall along a steep line, $L \propto M^{\alpha}$ , with $\alpha$ near $3.5$ – $4$ through the solar neighbourhood. Two main-sequence stars of the same mass are, to remarkable precision, the same star. What makes the relation deep rather than merely tidy is where it comes from. It is not a statement about nuclear physics. It falls out of the mechanical and radiative structure of a self-gravitating ball of gas: hydrostatic equilibrium fixes the internal pressure and temperature from mass and radius, and radiative diffusion sets the rate at which the trapped heat leaks out. The nuclear furnace merely adjusts its central temperature until the energy it generates matches this outward-diffusing luminosity — the luminosity is demanded by structure, not supplied by burning. This is why one can derive the MLR from the structure equations plus an opacity law, without ever writing down a reaction rate. The cleanest route is homology: assume two main-sequence stars have self-similar interiors, so that dimensionless profiles coincide once radius is scaled by $R$ and mass by $M$ . Feeding hydrostatic equilibrium, the radiative gradient, the ideal-gas law, and Kramers opacity into that assumption yields $L \propto \mu^{7.5}M^{5.5}R^{-0.5}$ ; because radius depends only weakly on mass, the residual $R^{-0.5}$ barely bends the exponent. This article takes the structure equations owned by its prerequisites and asks what luminosity they force — and where, at the extremes of mass, the simple power law must break.

Key takeaways

The mass–luminosity relation states that a main-sequence star's luminosity is a steep power of its mass, $L \propto M^{\alpha}$ , with $\alpha \approx 3.5$ – $4$ over most of the main sequence.
It follows almost entirely from the equations of stellar structure — hydrostatic equilibrium, the radiative gradient, and an opacity law — with no input from the nuclear energy-generation rate.
Under the homology assumption with Kramers opacity, $L \propto \mu^{7.5}M^{5.5}R^{-0.5}$ ; because $R$ varies weakly with $M$ , this collapses to a near-power law with $\alpha \approx 5$ , brought down toward the observed value by real (partly electron-scattering) opacities.
For massive stars electron scattering makes $\kappa$ constant and $L \propto \mu^{4}M^{3}$ ; near the top of the main sequence the luminosity approaches the Eddington limit $L_{\mathrm{Edd}} \propto M$ and the relation flattens to $\alpha \to 1$ .
The steepness explains stellar lifetimes: $\tau_{\mathrm{MS}} \sim M/L \propto M^{1-\alpha}$ plummets with mass, so a $10\,M_\odot$ star lives a few tens of Myr against the Sun's ten Gyr.
Key idea: For a main-sequence star luminosity is set by how fast energy diffuses out through the radiative envelope, which hydrostatic equilibrium ties directly to mass — so $L$ is a steep power of $M$ almost independent of the nuclear details.

Where this lives today

The mass–luminosity relation is working physics, not history. Detached eclipsing binaries observed by Gaia and ground-based surveys still provide the gold-standard calibration: each system with a measured orbit and eclipse depths yields two model-independent masses and luminosities, and the accumulated sample now pins the relation and its scatter to a few percent, tightening the stellar models that underlie almost everything else. Asteroseismology — reading pulsation frequencies from Kepler, TESS, and PLATO — delivers independent masses and radii that test the mass–radius input to the homology argument directly. On larger scales the relation is the hidden lever behind galactic mass-to-light ratios: because light is dominated by rare massive stars while mass hides in faint dwarfs, converting a galaxy's luminosity into a stellar mass requires the MLR folded through an initial mass function, and the mismatch with the dynamical mass from rotation curves is one of the original arguments for dark matter. The turnoff clock, powered by $\tau \propto M^{1-\alpha}$ , dates globular clusters and bounds the age of the Universe. And at the very top, the Eddington limit remains an active frontier: the physics that caps the MLR also governs the winds, eruptions, and pair-instability fate of the most massive stars — the progenitors of the black-hole mergers now heard by gravitational-wave detectors.

The physical picture

Picture a star as a leaky thermos. Gravity, through hydrostatic equilibrium, sets how hot and dense the interior must be to hold up a given mass. The luminosity is then just the rate at which that trapped heat leaks out through the insulating envelope — throttled by opacity, driven by the temperature gradient. A more massive star must be hotter inside to support its own weight, and hotter interiors leak radiation dramatically faster, so massive stars are enormously more luminous than their extra mass alone would suggest. The nuclear furnace is a thermostat, not a throttle. If burning ran faster than the leak rate the core would heat, expand, and cool, quenching the reactions; if slower, it would contract and heat until it caught up. So the outward radiative leak — pure structure — dictates $L$ , and the reactions silently match it. That decoupling is exactly why the MLR can be derived without a single reaction rate, and the first figure traces the causal chain from mass to luminosity.

Figure

The causal chain behind the MLR: mass sets the interior via hydrostatic equilibrium and the ideal-gas law to give T_c ~ μM/R (about 1.5×10⁷ K for the Sun), radiative diffusion then leaks the luminosity, and with Kramers opacity the chain closes on L ∝ μ^7.5 M^5.5 / R^0.5. Nuclear burning is a thermostat that matches this demanded L, it does not set it

Because $L$ climbs so much faster than $M$ , the fuel-to-burn-rate ratio $M/L$ collapses with mass: a $10\,M_\odot$ star carries ten times the Sun's fuel but shines thousands of times brighter, so it burns out in tens of millions of years while the Sun lasts ten billion. And at the very top, photons push outward nearly as hard as gravity pulls in (the Eddington limit), so luminosity can no longer rise steeply and the relation bends toward $L \propto M$ . The second figure is the relation itself, across four decades of mass, with those regime changes marked.

Figure

The mass–luminosity relation on log–log axes: the measured main sequence runs steeply as roughly L ∝ M^3.5 through the solar point (1,1), shallows to α ≈ 2.3 for red dwarfs below 0.4 M⊙, and flattens toward α ≈ 1 at the top where it presses against the Eddington ceiling L_Edd = 3.8×10⁴ (M/M⊙) L⊙ shown as the dashed line

The slope changes across the diagram because the physics that carries the heat changes: fully convective red dwarfs with molecular and hydrogen-negative-ion opacities shallow it at the bottom, a radiative Kramers envelope gives the steep solar section, and higher up electron scattering eases the exponent until radiation pressure caps the luminosity flat against the Eddington ceiling. One diagram, three regimes — gravity, thermodynamics, and photon transport, exactly what the structure equations describe.

The core equations

The relation begins as an empirical fact and becomes a theorem. We state the observed power law and its calibration, then the four structure ingredients that generate it, then the two homology results and the Eddington ceiling they run into. An equals sign marks an exact relation; a scaling symbol ( $\propto$ or $\sim$ ) marks a homology or order-of-magnitude estimate with dimensionless constants dropped. The headline relation is the observed one: luminosity is a steep power of mass.

L \propto M^{\alpha}, \qquad \alpha \approx 3.5\ \text{(main sequence)} \qquad (1)

where: $L$ — total (bolometric) luminosity; $M$ — stellar mass; $\alpha$ — logarithmic slope, running from $\sim 1$ (most massive) through $\sim 4$ (solar) to $\sim 2.3$ (lowest mass). Reading: near the Sun a factor of two in mass is roughly an order of magnitude in luminosity — mass is the master variable of a main-sequence star. Calibrated on eclipsing and visual binaries — the only stars with directly weighed masses — the local relation is often quoted logarithmically near one solar mass.

\log_{10}\!\left(\frac{L}{L_\odot}\right) \approx 3.8\,\log_{10}\!\left(\frac{M}{M_\odot}\right) \qquad (2)

where: $L_\odot = 3.828\times10^{26}\,\mathrm{W}$ — solar luminosity; $M_\odot = 1.989\times10^{30}\,\mathrm{kg}$ — solar mass. Reading: the slope on a log $L$ –log $M$ plot is $\alpha$ — near the Sun about $3.8$ , and a local slope, not a universal constant. The first structural ingredient is hydrostatic equilibrium: at every layer the pressure gradient supports the weight of the overlying gas.

\frac{\mathrm{d}P}{\mathrm{d}r} = -\frac{G\,m(r)\,\rho}{r^{2}} \qquad (3)

where: $P$ — pressure at radius $r$ ; $m(r)$ — mass enclosed within $r$ ; $\rho$ — local density; $G$ — gravitational constant. Reading: integrated across the star this fixes the central pressure scale $P_c \sim GM^{2}/R^{4}$ — more mass in less radius means higher pressure. The second ingredient is the radiative temperature gradient — Fick's law for photons — which sets how steeply temperature must fall to push a luminosity $L(r)$ outward against the opacity.

\frac{\mathrm{d}T}{\mathrm{d}r} = -\frac{3}{4ac}\,\frac{\kappa \rho}{T^{3}}\,\frac{L(r)}{4\pi r^{2}} \qquad (4)

where: $T$ — temperature; $\kappa$ — Rosseland-mean opacity per unit mass; $a = 4\sigma/c$ — radiation constant; $c$ — speed of light; $L(r)$ — luminosity crossing radius $r$ . Reading: solved for the flux this is the leak law, $L \sim ac\,R\,T_c^{4}/(\kappa\rho)$ — the equation that turns interior temperature into radiated power. The third ingredient is the equation of state: for an ideal gas the pressure ties temperature to density and composition.

P = \frac{\rho\,k_{B}\,T}{\mu\,m_{H}} \qquad (5)

where: $k_B$ — Boltzmann constant; $\mu$ — mean molecular weight (mean particle mass in units of $m_H$ ); $m_H$ — hydrogen mass. Reading: combined with (3) it gives $T_c \sim \mu m_H GM/(k_B R)$ — hotter for larger $M/R$ and heavier composition $\mu$ . The fourth ingredient is opacity: across much of a Sun-like interior the dominant bound-free and free-free absorption follows the Kramers law.

\kappa = \kappa_0\,\rho\,T^{-3.5} \qquad (6)

where: $\kappa_0$ — composition-dependent constant. Reading: opacity rises with density and plummets with temperature — hotter gas is more transparent — and it is this steep $T^{-3.5}$ that injects the large exponents into the MLR. (Free–free absorption per ion scales as $\nu^{-3}T^{-1/2}$ ; Rosseland-averaging over a thermal spectrum adds $T^{-3}$ , and the absorber density $\propto \rho$ gives the $\rho\,T^{-3.5}$ form.) Assembling (3)–(6) under homology gives the classic Kramers mass–luminosity relation, derived in full below.

L \propto \frac{\mu^{7.5}\,M^{5.5}}{\kappa_0\,R^{0.5}} \qquad (7)

where: $R$ — stellar radius; $\mu$ — mean molecular weight. Reading: luminosity is enormously sensitive to composition ( $\mu^{7.5}$ ) and mass ( $M^{5.5}$ ) but weak in radius. Since $R \propto M^{0.6\text{–}0.8}$ , the $R^{-0.5}$ term barely lowers the exponent, leaving $\alpha \approx 5$ for pure Kramers opacity. For hot, massive interiors electron scattering dominates, $\kappa$ becomes constant, and the radius dependence cancels exactly.

L \propto \mu^{4}\,M^{3} \qquad (8)

where: $\mu$ — mean molecular weight. Reading: a cleaner power law with no $R$ dependence, giving $\alpha \approx 3$ — the regime governing luminous, intermediate-to-high-mass stars. The ceiling on all of this is the Eddington luminosity, where radiation pressure alone would support the outer layers against gravity.

L_{\mathrm{Edd}} = \frac{4\pi G M c}{\kappa_{es}} \approx 3.8\times10^{4}\,\frac{M}{M_\odot}\,L_\odot \qquad (9)

where: $\kappa_{es} \approx 0.034\,\mathrm{m^2\,kg^{-1}}$ — electron-scattering opacity for solar composition; $L_{\mathrm{Edd}}$ — maximum luminosity before radiation unbinds the envelope. Reading: strictly linear in $M$ and independent of radius and composition, so a star radiating near it has its slope forced toward $\alpha = 1$ . Finally, the steep MLR sets the clock of stellar evolution through the fuel-to-power ratio.

\tau_{\mathrm{MS}} \sim \frac{E_{\mathrm{nuc}}}{L} \propto \frac{M}{L} \propto M^{\,1-\alpha} \qquad (10)

where: $\tau_{\mathrm{MS}}$ — main-sequence lifetime; $E_{\mathrm{nuc}} \propto M$ — nuclear energy reservoir. Reading: with $\alpha \approx 3.5$ the lifetime scales as $M^{-2.5}$ — massive stars are spendthrifts, low-mass stars misers — the basis of the turnoff clock used to age clusters.

Derivation AThe full homology mass–luminosity relation

The goal is to show that the four structure equations alone — hydrostatic equilibrium, the ideal-gas law, Kramers opacity, and the radiative gradient — force $L \propto \mu^{7.5}M^{5.5}R^{-0.5}$ without any reference to nuclear energy generation, and then to reduce this to a near-power law in mass alone using the observed mass–radius relation. The tool is homology: we assume two main-sequence stars are self-similar, so that when we scale radius by $R$ and mass by $M$ their dimensionless profiles $P(r/R)$ , $T(r/R)$ , $\rho(r/R)$ , and $m(r)/M$ coincide. Under that assumption every physical variable at a corresponding fractional radius scales as a fixed power of the global $M$ and $R$ , and the derivatives in the structure equations can be replaced by ratios of characteristic scales.

Step 1. Take hydrostatic equilibrium, $\mathrm{d}P/\mathrm{d}r = -Gm\rho/r^{2}$ , and replace derivatives by scales: $P/R \sim GM\rho/R^{2}$ with $\rho \sim M/R^{3}$ , giving $P \propto M^{2}/R^{4}$ . Why: the mechanical backbone — pressure is set by how much mass is packed into how small a radius, before any thermal physics.

Step 2. Insert this into the ideal-gas law $P = \rho k_B T/(\mu m_H)$ : $T \propto \mu P/\rho \propto \mu (M^{2}/R^{4})/(M/R^{3}) = \mu M/R$ . Why: support plus an ideal-gas law means the interior temperature is not free — a bigger, more compact, or heavier-composition star runs hotter.

Step 3. Feed the density and temperature scalings into Kramers opacity, $\kappa \propto \kappa_0\,(M/R^{3})(\mu M/R)^{-3.5}$ . Why: opacity converts the thermodynamic state into a resistance photons must diffuse against, and its steep temperature sensitivity inflates the mass exponent.

Step 4. Solve the radiative gradient for the flux: with $\mathrm{d}T/\mathrm{d}r \sim T/R$ and $r \sim R$ , equation (4) becomes $L \propto T^{4}R/(\kappa\rho)$ . Why: the leak law in scaling form — the diffusible luminosity is the radiative energy density $\propto T^4$ carried across the star over resistance $\kappa\rho$ .

Step 5. Substitute and collect exponents: $L \propto (\mu M/R)^{4}\,R \big/ \big[\kappa_0(M/R^{3})(\mu M/R)^{-3.5}(M/R^{3})\big]$ , i.e. numerator $\mu^{4}M^{4}R^{-3}$ over denominator $\kappa_0\,\mu^{-3.5}M^{-1.5}R^{-2.5}$ . Why: pure bookkeeping, but this is where the large exponents are assembled — every power of $M$ and $R$ must be tracked exactly.

Step 6. Divide to obtain $L \propto \mu^{7.5}M^{5.5}R^{-0.5}/\kappa_0$ . Why: the steep dependence on mass and composition is the origin of the MLR — and note the nuclear rate never appeared, so luminosity is structural, not nuclear.

Step 7. Reduce to $M$ alone. With the main-sequence mass–radius relation $R \propto M^{0.6\text{–}0.8}$ , $L \propto M^{\,5.5 - 0.5\times(0.6\text{–}0.8)} \approx M^{5.1\text{–}5.2}$ . Why: the weak $R^{-0.5}$ term barely dents the exponent, so pure Kramers homology predicts $\alpha \approx 5$ ; real opacities (partly electron scattering) and non-homology bring the observed value down to $3.5$ – $4$ . *The exponent bookkeeping gives $\mu^{7.5}M^{5.5}R^{-0.5}$ exactly: numerator $\mu^4 M^4 R^{-3}$ over denominator $\mu^{-3.5}M^{-1.5}R^{-2.5}$ .*

Go deeper · Derivation B · 2 minThe electron-scattering limit and the Eddington ceiling

Derivation BThe electron-scattering limit and the Eddington ceiling

The goal is to explain why the steep Kramers relation must give way at high mass to a shallower law and ultimately to a hard ceiling. Two things change as we climb the main sequence: the opacity source switches from Kramers absorption to electron scattering, which is independent of temperature and density, and radiation pressure grows until it rivals gravity. We take the same homology machinery, swap the opacity law, and then balance radiative and gravitational forces in the outer envelope to find the maximum luminosity a star of mass $M$ can sustain — the point at which the mass–luminosity relation is forced flat.

Step 1. Replace Kramers opacity with a constant $\kappa = \kappa_{es}$ in the leak law $L \propto T^{4}R/(\kappa\rho)$ , giving $L \propto R(\mu M/R)^{4}/(M/R^{3})$ . Why: for hot ionized interiors free electrons dominate the scattering and $\kappa$ stops depending on local state, removing the temperature sensitivity that steepened the Kramers exponent.

Step 2. Cancel the radius factors: $R \cdot R^{-4} \cdot R^{3} = R^{0}$ , so $L \propto \mu^{4}M^{3}$ . Why: with a constant opacity the star's size drops out, leaving the cleaner power law $\alpha \approx 3$ that governs the near- $M^3$ section of the observed relation.

Step 3. Now balance forces in the outer layers: each gram of opacity $\kappa_{es}$ feels an outward $g_{\mathrm{rad}} = (\kappa_{es}/c)\,L/(4\pi r^{2})$ against inward gravity $g = GM/r^{2}$ . Why: radiation carries momentum and an ionized envelope absorbs it efficiently, so a bright enough star can lift its own atmosphere.

Step 4. Set $g_{\mathrm{rad}} = g$ ; the $r^{-2}$ factors cancel, giving $L_{\mathrm{Edd}} = 4\pi G M c/\kappa_{es}$ , linear in $M$ and independent of radius and composition. Why: the Eddington luminosity, an upper bound every star must respect — exceeding it means net outward force and mass loss.

Step 5. Combine with $L \propto M^{3}$ : as mass grows the luminosity rises far faster than the linear ceiling, driving the star toward $L_{\mathrm{Edd}}$ and bending the slope from $3$ down to $L \to L_{\mathrm{Edd}} \propto M$ , i.e. $\alpha \to 1$ . Why: this flattens the top of the main sequence and, through winds and instabilities near $L_{\mathrm{Edd}}$ , sets the practical upper mass limit of $\sim 100$ – $150\,M_\odot$ . *The force balance gives $L_{\mathrm{Edd}} = 4\pi GMc/\kappa_{es} = 3.8\times10^{4}(M/M_\odot)L_\odot$ for $\kappa_{es} = 0.034\,\mathrm{m^2\,kg^{-1}}$ ; the electron-scattering homology gives $L \propto \mu^4 M^3$ with radius cancelling exactly.*

Go deeper · Derivation C · 1 minThe main-sequence lifetime and the stellar clock

Derivation CThe main-sequence lifetime and the stellar clock

The goal is to turn the mass–luminosity relation into a statement about time — to show why massive stars die young and low-mass stars effectively never leave the main sequence, and to make the turnoff of a star cluster into a usable age. The nuclear fuel available to a star is a fixed fraction of its rest energy and scales with mass, while the rate at which that fuel is spent is the luminosity; dividing the two gives the burning time, and substituting the MLR converts a lifetime into a single power of mass.

Step 1. Write the reservoir: a star fuses a roughly fixed fraction of its mass at efficiency $\sim 0.007c^{2}$ per unit mass, so $E_{\mathrm{nuc}} \propto M$ . Why: to leading order only mass sets the fuel budget — a heavier star has proportionally more hydrogen.

Step 2. Divide reservoir by power: $\tau_{\mathrm{MS}} \sim E_{\mathrm{nuc}}/L \propto M/L$ . Why: burning time is fuel over burn-rate, and since burn-rate is the structurally-demanded luminosity, the lifetime inherits the MLR directly.

Step 3. Substitute $L \propto M^{\alpha}$ to get $\tau_{\mathrm{MS}} \propto M^{\,1-\alpha}$ ; with $\alpha \approx 3.5$ , $\tau \propto M^{-2.5}$ . Why: the steepness of the MLR is exactly what makes lifetime so violently mass-dependent — extra fuel is overwhelmed by far larger luminosity.

Step 4. Anchor to $\tau_\odot \approx 10^{10}$ yr: a $10\,M_\odot$ star lives $10^{10}\times10^{-2.5} \approx 3\times10^{7}$ yr, a $0.2\,M_\odot$ dwarf $\gg 10^{11}$ yr — longer than the age of the Universe. Why: the basis of the turnoff clock — in a coeval cluster the most massive surviving stars pinpoint the age, everything more massive having evolved off. *As a check: for $\alpha = 3.5$ , $16^{-2.5} = 1/1024 \approx 10^{-3}$ and $10^{1-3.5} = 3.2\times10^{-3}$ ; anchoring to $\tau_\odot = 10^{10}$ yr then reproduces the cluster-age scalings.*

Limiting cases

The exponent $\alpha$ in $L \propto M^{\alpha}$ is not a constant: it tracks whichever opacity and equation of state dominate the interior, and those change systematically along the main sequence.

Case	Result	Meaning
Low mass ( $M \lesssim 0.4\,M_\odot$ )	$L \propto M^{2.3}$	Fully convective interiors with hydrogen-negative-ion and molecular opacities; the relation is shallower than the classic value, and lifetimes exceed the age of the Universe.
Solar regime ( $0.4 \lesssim M/M_\odot \lesssim 2$ )	$L \propto M^{4}$	Radiative envelope with Kramers-like opacity; the steepest observed section, $\alpha \approx 3.5$ – $4$ , where the homology derivation applies most directly.
Intermediate ( $2 \lesssim M/M_\odot \lesssim 20$ )	$L \propto M^{3.5}$	Electron scattering begins to compete with Kramers; the exponent eases toward the $\mu^{4}M^{3}$ value.
Very massive ( $M \gtrsim 20\,M_\odot$ )	$L \propto M$	Radiation pressure and electron scattering dominate; luminosity presses against $L_{\mathrm{Edd}} \propto M$ and the slope flattens to $\alpha \to 1$ .

Common trapThe MLR is nuclear physics — and α is a universal constant

Both halves of this are wrong, and they are the two most common misconceptions about the relation. First, the mass–luminosity relation is not caused by nuclear burning. The luminosity is set by how fast energy diffuses out through the radiative envelope — pure hydrostatic-plus-radiative structure — and the burning rate self-adjusts through the thermostat effect to supply exactly that. This is why the derivation above never invokes an energy-generation rate: change the nuclear physics and the star simply finds a slightly different central temperature at the same $L$ . Second, $\alpha$ is not a single number. It slides from about $2.3$ for red dwarfs, through $\approx 4$ near the Sun, down toward $1$ for the most massive stars, because the dominant opacity and equation of state change with mass. Quoting " $L \propto M^{3.5}$ " without naming a mass range is a slogan, not a law — always attach the regime.

Worked examples

1 · The Eddington ceiling and how close a massive star runs to it (fully worked)

Compute the Eddington luminosity per solar mass for solar composition, then evaluate it for a $60\,M_\odot$ O-type star and compare with that star's actual luminosity of about $6\times10^{5}\,L_\odot$ . What does the ratio tell you?

Step 1. Assemble the constants: $G = 6.674\times10^{-11}\,\mathrm{m^3\,kg^{-1}\,s^{-2}}$ , $c = 2.998\times10^{8}\,\mathrm{m\,s^{-1}}$ , $M_\odot = 1.989\times10^{30}\,\mathrm{kg}$ , and the solar-composition electron-scattering opacity $\kappa_{es} = 0.2(1+X) = 0.34\,\mathrm{cm^2\,g^{-1}} = 0.034\,\mathrm{m^2\,kg^{-1}}$ with $X = 0.7$ .

Step 2. Evaluate $L_{\mathrm{Edd}} = 4\pi G M_\odot c/\kappa_{es}$ . The numerator is $4\pi \times 6.674\times10^{-11} \times 1.989\times10^{30} \times 2.998\times10^{8} = 5.00\times10^{29}$ , and dividing by $0.034$ gives $\mathbf{L_{Edd} = 1.47\times10^{31}\,W}$ per solar mass. In solar units, $1.47\times10^{31}/3.828\times10^{26} = \mathbf{3.8\times10^{4}\,L_\odot}$ per $M_\odot$ .

Step 3. Scale to the star: $L_{\mathrm{Edd}}(60\,M_\odot) = 60 \times 3.8\times10^{4} = \mathbf{2.3\times10^{6}\,L_\odot}$ .

Step 4. Form the Eddington factor $\Gamma = L/L_{\mathrm{Edd}} = 6\times10^{5}/2.3\times10^{6} = \mathbf{0.26}$ . The star already radiates at about a quarter of the luminosity that would blow its atmosphere off — which is why O-stars drive powerful radiation-line-driven winds, and why anything much more massive would be pushed over the limit. The Eddington ceiling is 3.8×10⁴ (M/M⊙) L⊙, and a 60 M⊙ star runs at Γ ≈ 0.26 of it.

2 · From mass to luminosity to lifetime (fully worked)

A detached eclipsing binary yields a main-sequence component of $10\,M_\odot$ . Estimate its luminosity and its main-sequence lifetime, and compare with the Sun.

Step 1. Choose the regime. At $10\,M_\odot$ we are in the intermediate range, so use $L \propto M^{3.5}$ : $L = (10)^{3.5}\,L_\odot = \mathbf{3.2\times10^{3}\,L_\odot}$ . The star is three thousand times more luminous than the Sun on ten times the mass — the signature steepness of the MLR.

Step 2. Apply the lifetime scaling $\tau_{\mathrm{MS}} \propto M^{1-\alpha}$ anchored to $\tau_\odot \approx 10^{10}$ yr: $\tau = 10^{10}\times(10)^{1-3.5} = 10^{10}\times10^{-2.5} = \mathbf{3.2\times10^{7}\,yr}$ , about 30 million years.

Step 3. The accounting: ten times the fuel burned three thousand times faster lasts about $10/3200 \approx 1/300$ as long — tens of Myr against ten Gyr. A 10 M⊙ star shines at ≈ 3200 L⊙ and lives ≈ 30 Myr, about 300× shorter than the Sun.

3 · Why red dwarfs are immortal

Estimate the luminosity and lifetime of a $0.2\,M_\odot$ red dwarf.

Step 1. Use the low-mass slope $L \propto M^{2.3}$ : $L = (0.2)^{2.3}\,L_\odot = \mathbf{0.025\,L_\odot}$ — about one-fortieth of the Sun.

Step 2. Lifetime: $\tau = 10^{10}\times(M/L) = 10^{10}\times(0.2/0.025) = 10^{10}\times 8 = \mathbf{8\times10^{10}\,yr}$ — and this is a floor, because low-mass stars are fully convective and can burn nearly all their hydrogen, stretching the true lifetime to $\sim10^{12}$ yr.

Step 3. The point: the same steep relation that dooms massive stars makes the faintest ones effectively eternal. No red dwarf has ever left the main sequence in cosmic history, which is why they will dominate the far future of the Galaxy long after every Sun-like star has died. A 0.2 M⊙ dwarf shines at ≈ 0.025 L⊙ and outlives the present Universe many times over.

Physics of the portrait

The mass–luminosity relation is where a measured empirical curve became a theorem of physics, and one figure did both the theory and the naming: Arthur Eddington, who derived it from the structure of a radiating gas sphere and, in the same work, found the luminosity ceiling that bears his name.

Origins · 1 minArthur Eddington and how this was discovered

Arthur Eddington

1882–1944 · the theory of stellar structure and radiative equilibrium

Arthur Stanley Eddington was the astrophysicist who first understood a star as a self-regulating ball of hot gas held up against gravity by the pressure of its own radiation and matter. Through the early 1920s, culminating in his 1926 masterpiece The Internal Constitution of the Stars, he built the "standard model" of a star from hydrostatic equilibrium, an ideal-gas-plus-radiation pressure law, and radiative energy transport — the very equations this article assembles. When he plotted the theoretical luminosities against mass he found the steep dependence observers had begun to notice in binaries, and in his 1924 paper he announced the mass–luminosity relation as a prediction of the theory, not a fit to data. Strikingly, he obtained it before the source of stellar energy was known: whatever the source, he argued, the luminosity is fixed by how fast radiation diffuses out, so the burning must adjust to match — the thermostat argument in modern form. The same analysis exposed the radiation-pressure limit now called the Eddington luminosity, $L_{\mathrm{Edd}} = 4\pi GMc/\kappa$ , above which a star cannot hold its outer layers, and with it the reason the most massive stars are rare and unstable. Eddington's insistence that structure, not nuclear detail, sets a star's luminosity — contentious in his day, and the root of his famous clashes with Jeans and Chandrasekhar — is the foundation on which the entire field of stellar structure now rests.

History

1911

— Hertzsprung and Russell chart luminosity against spectral type, revealing the main sequence

1919

— Eddington models a star as a gas sphere in radiative equilibrium with significant radiation pressure

1924

— Eddington publishes the theoretical mass–luminosity relation and the radiation-pressure luminosity limit

1926

— The Internal Constitution of the Stars sets out the standard model of a radiative star

1938

— Bethe identifies the pp-chain and CNO reactions that silently supply the structurally-demanded luminosity

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 10–11 (stellar structure, homology, and the mass–luminosity relation) · Kippenhahn, Weigert & Weiss, Stellar Structure and Evolution, 2nd ed., Ch. 20 (homology and the M–L relation) · Hansen, Kawaler & Trimble, Stellar Interiors, 2nd ed. · Prialnik, An Introduction to the Theory of Stellar Structure and Evolution, 2nd ed. · Mo, van den Bosch & White, Galaxy Formation and Evolution, §10 (stellar populations and mass-to-light ratios). The key numerical results are: (1) Kramers-homology bookkeeping gives $L \propto \mu^{7.5}M^{5.5}R^{-0.5}$ (numerator $\mu^4 M^4 R^{-3}$ over denominator $\kappa_0\,\mu^{-3.5}M^{-1.5}R^{-2.5}$ ); (2) electron-scattering homology gives $L \propto \mu^{4}M^{3}$ with radius cancelling; (3) $L_{\mathrm{Edd}} = 4\pi GMc/\kappa_{es} = 3.8\times10^{4}(M/M_\odot)L_\odot$ with $\kappa_{es} = 0.034\,\mathrm{m^2\,kg^{-1}}$ ; (4) $\tau \propto M^{1-\alpha}$ with $16^{-2.5} = 1/1024$ . Constants used throughout: $L_\odot = 3.828\times10^{26}\,\mathrm{W}$ , $M_\odot = 1.989\times10^{30}\,\mathrm{kg}$ , $G = 6.674\times10^{-11}$ SI, $c = 2.998\times10^{8}\,\mathrm{m\,s^{-1}}$ .

Prerequisites. Hydrostatic Equilibrium in Stars · Equations of Stellar Structure · Radiative Temperature Gradient · Opacity and Radiative Transfer in Stars · Equation of State for Stellar Matter. Leads to. Radiation Pressure in Stellar Interiors · Convective Energy Transport · Energy Generation Equation.

Ask the tutor: Ask me to weigh any star from its luminosity, walk through the homology bookkeeping line by line, or show how the Eddington limit caps the top of the main sequence.

Is the mass–luminosity relation an observed fact or a theoretical prediction?

It is both, and that is what makes it powerful. Observationally it was established from binary stars — detached eclipsing and visual binaries are the only systems where masses are weighed directly through Kepler's laws, and their luminosities plotted against those masses trace a steep line. Theoretically, Eddington showed the same relation falls out of the equations of stellar structure without any fit. The agreement between a curve measured from orbital mechanics and one derived from hydrostatic and radiative physics is among the cleanest confirmations that we understand what holds a star up and how it leaks its heat. Where the two disagree, at the extremes of mass, the disagreement itself flags where convection, degeneracy, or radiation pressure has taken over.

If luminosity is set by structure, what does the nuclear burning actually do?

It keeps the star from collapsing, at whatever rate the structure demands. The luminosity is fixed by how fast radiation diffuses out of the envelope, and that leak would drain the star's heat and let it contract if nothing replenished it. Burning is the replenishment: the core settles at exactly the temperature where energy generated equals energy leaking out. Overshoot, and the core heats, expands, and cools, throttling the reactions back; undershoot, and it contracts and heats until it catches up. So burning sets the star's lifetime (how long the fuel lasts) but not its instantaneous luminosity — which is why the MLR can be derived without ever writing a reaction rate.

Why is the homology exponent about 5 when the observed one is 3.5 to 4?

Because pure Kramers opacity is only part of the story. Kramers homology gives $L \propto M^{5.5}R^{-0.5}$ , which the weak mass–radius relation trims only to $\alpha \approx 5$ . But real interiors are not governed by Kramers alone — electron scattering, which gives the shallower $L \propto M^{3}$ , contributes increasingly at higher mass, and the true opacity is a blend. Real stars are also not perfectly homologous: central concentration, convective cores, and composition gradients all vary with mass. Together these pull the effective exponent down to the observed $3.5$ – $4$ . The homology value is best read as an upper bracket that isolates the physics, not a precise prediction.

What sets the mean molecular weight μ, and why does the relation depend on it so strongly?

The mean molecular weight $\mu$ is the average particle mass in units of the hydrogen mass: fully ionized hydrogen gives $\mu \approx 0.5$ , fully ionized helium about $1.33$ , and solar-composition ionized gas about $0.62$ . It enters the ideal-gas law, so a higher $\mu$ means fewer, heavier particles must supply the same pressure, forcing a hotter interior — and because luminosity depends on interior temperature so steeply, the Kramers result carries a $\mu^{7.5}$ factor. This matters physically: as a star fuses hydrogen to helium its core $\mu$ rises, the interior heats, and the luminosity slowly climbs even on the main sequence, which is why the Sun is about 30% brighter now than at its birth.

Does the mass–luminosity relation apply to giants and white dwarfs?

No — it is specifically a main-sequence relation, valid only in the core-hydrogen-burning phase, and this is a common trap. The derivation assumes an ideal gas in radiative equilibrium with a normal opacity, which describes main-sequence stars. A red giant has a tiny, near-degenerate core and a vastly extended convective envelope; its luminosity is set by shell burning and core mass, not total mass, so it lies far above the line. A white dwarf has no fusion at all — a cooling degenerate remnant whose luminosity depends on age and residual heat, with an inverse mass–radius relation ( $R$ shrinks as $M$ grows) that breaks every homology assumption. Restrict the MLR to the main sequence.

How does the steep relation lead to an upper limit on stellar mass?

Through the Eddington limit and the instabilities near it. Because $L$ grows faster than $M$ while the ceiling $L_{\mathrm{Edd}} \propto M$ grows only linearly, a more massive star sits ever closer to the point where radiation pressure overwhelms gravity in its outer layers. Approaching $L_{\mathrm{Edd}}$ , its radiation-driven wind strengthens sharply and it becomes prone to pulsational and continuum-driven instabilities that shed mass in violent episodes — as in luminous blue variables like Eta Carinae. These cap the mass a star can retain at roughly $100$ – $150\,M_\odot$ , with the exact ceiling depending on metallicity through wind strength. The same relation that governs ordinary stars, pushed to its limit, is why supermassive stars do not exist.