Stefan-Boltzmann Law

A power law that nature actually obeys

Every object above absolute zero radiates electromagnetic energy. Cold objects radiate weakly and at long wavelengths; hot objects radiate strongly and at shorter wavelengths. The simplest possible question — how does the total radiated power depend on temperature? — has an unusually clean answer: it scales as the fourth power of the absolute temperature.

P / A = σ T⁴            (per unit area, into a hemisphere)

σ = 5.670374419 × 10⁻⁸ W/(m²·K⁴)   Stefan-Boltzmann constant
T  = absolute temperature in kelvin

The fourth-power dependence is steep. A small change in T multiplies the radiated power dramatically. A factor of 2 in T multiplies P by 16; a factor of 10 multiplies P by 10000. This is why high-temperature processes (combustion, fusion, industrial furnaces, stellar atmospheres) are dominated by radiative rather than conductive heat transfer, and why infrared cameras can pick out a person against a 20°C wall: human skin at 305 K radiates 1.07× the wall flux at 293 K — a 7 % contrast that the camera resolves easily.

The constant σ is one of the half-dozen most important numbers in physics. Its modern definition is exact, derived from other defined constants: σ = 2π⁵k_B⁴/(15h³c²), where k_B is Boltzmann's constant, h is Planck's constant and c is the speed of light. Knowing σ to 11 significant figures is what lets us convert star temperatures to luminosities, kiln temperatures to power inputs and thermograph signals to surface temperatures.

Stefan, Boltzmann, and the road to the law

In 1879 Josef Stefan, working in Vienna, fit experimental data on heated platinum filaments collected by John Tyndall in London a decade earlier. Stefan noticed that radiated power went up roughly as T⁴, and proposed the law in its modern form. The fit was good but not perfect — Tyndall's measurements suffered from convective losses that masked the pure radiative term — and a generation of physicists took the law more as suggestive than proven.

Ludwig Boltzmann, then teaching at Graz, settled the matter in 1884 with a thermodynamic derivation. He treated radiation in a cavity as a working fluid in a Carnot cycle, using Maxwell's prediction that electromagnetic radiation exerts pressure P = u/3 (where u is energy density). Applying ∂u/∂T thermodynamic identities to a hypothetical engine running on radiation gave u ∝ T⁴. Total energy emitted per area follows because radiation in the cavity has speed c and isotropic distribution: emission rate is u·c/4. The numerical constant Boltzmann obtained matched Stefan's empirical value within experimental error.

The deepest derivation came in 1900, when Max Planck found the spectral distribution of cavity radiation by quantising the oscillator energies. Integrating Planck's spectral formula over all wavelengths reproduces Stefan-Boltzmann exactly. The integration constant becomes σ = 2π⁵k_B⁴/(15h³c²) — the first appearance of h in any practical formula. In a sense Stefan-Boltzmann was the first quantum law, even though the physicists involved hadn't realised it.

Worked example: the Sun's surface flux and luminosity

The Sun's photospheric effective temperature is T_⊙ = 5778 K. To compute the radiated flux at the surface:

P / A = σ T⁴
      = 5.670 × 10⁻⁸ × (5778)⁴
      = 5.670 × 10⁻⁸ × 1.1148 × 10¹⁵
      = 6.32 × 10⁷ W/m²
      ≈ 63 MW/m²

Sixty-three megawatts per square metre. A commercial nuclear reactor produces roughly 3000 MW, the entire output filling a city block; the equivalent solar surface area is just 50 m² — about the area of a small living room. Now multiply by the Sun's total surface area:

A_⊙ = 4π R_⊙² = 4π × (6.96 × 10⁸)² = 6.087 × 10¹⁸ m²

L_⊙ = (P/A) × A_⊙
    = 6.32 × 10⁷ × 6.087 × 10¹⁸
    = 3.85 × 10²⁶ W

That is the solar luminosity, the total radiative power output of the Sun. The accepted value to four digits is 3.828 × 10²⁶ W; our Stefan-Boltzmann estimate matches to better than 1 %. The same recipe — measure the effective temperature, apply σT⁴, multiply by surface area — gives the luminosity of every star whose temperature can be obtained from its spectrum.

Worked example: Earth's effective temperature

Earth absorbs sunlight on its illuminated cross-section πR² and emits thermal radiation from its full surface 4πR². In radiative equilibrium, absorbed power equals emitted power. The solar constant at Earth's orbit is S = 1361 W/m² and Earth's mean albedo (reflectivity) is A = 0.30:

Absorbed: (1 − A) × S × π R²
Emitted:   σ T_eff⁴ × 4 π R²

(1 − A) × S = 4 σ T_eff⁴

T_eff = [(1 − A) × S / (4 σ)]^(1/4)
      = [0.70 × 1361 / (4 × 5.670 × 10⁻⁸)]^(1/4)
      = [952.7 / 2.268 × 10⁻⁷]^(1/4)
      = [4.20 × 10⁹]^(1/4)
      = 254.6 K
      ≈ −18.5°C

Earth's effective radiating temperature is about −18°C. The actual mean surface temperature is +15°C, 33 K higher. The 33 K gap is the greenhouse effect — surface infrared photons are absorbed by water vapour, CO₂ and methane at altitudes a few kilometres up, and re-emitted in all directions, with about half going back down. The surface ends up warmer than the radiating layer by exactly the amount needed for outgoing radiation from the top of the atmosphere to balance incoming sunlight.

This same calculation, repeated for every planet, predicts equilibrium temperatures within a few kelvin once you have its albedo and orbit:

Planet/object	Solar flux S (W/m²)	Albedo A	T_eff (K)	Actual surface T (K)
Mercury	9083	0.12	434	440 (mean)
Venus	2611	0.76	227	737 (CO₂ greenhouse)
Earth	1361	0.30	255	288
Mars	586	0.25	210	210
Jupiter (cloud tops)	50	0.34	110	110
Pluto	0.87	0.50	37	44

The Stefan-Boltzmann prediction is nearly exact for Mars and Jupiter (no significant atmosphere or insulating greenhouse), modestly off for Earth (greenhouse warming), and dramatically off for Venus, where the dense CO₂ atmosphere boosts surface temperature by 510 K above radiative equilibrium. Climate-relevant CO₂ physics is essentially a problem in non-grey radiative transfer riding on top of the σT⁴ baseline.

Real surfaces and emissivity

A perfect blackbody radiates exactly σT⁴ per unit area. Real surfaces radiate less. The fraction of blackbody emission a real surface achieves is its emissivity ε(T,λ,direction), a number between 0 and 1. For "grey body" approximations where ε is taken constant in wavelength, the law becomes:

P / A = ε σ T⁴

Materials span the full range. Polished metals have very low emissivity; oxidised, painted or rough surfaces have high emissivity. This is why thermos flasks have silvered inner walls and why aluminised plastic emergency blankets reduce radiative heat loss from the human body by 90 %.

Surface	Emissivity ε	Note
Polished silver	0.02	Used in radiation shields
Polished aluminium	0.04	Mylar foil, mirror coatings
Oxidised aluminium	0.15	Common roofing
Stainless steel (brushed)	0.40	Industrial
Human skin	0.98	Why IR thermometers work
Soot, lampblack	0.95	Reference for "matt black"
Water (still surface)	0.96	Effectively black at IR
Snow, fresh	0.99	High at thermal IR

By Kirchhoff's law of thermal radiation, at any wavelength a surface's emissivity equals its absorptivity. A good emitter is a good absorber and vice versa. This is why polished aluminium is used both inside and outside thermos flasks — it absorbs little outside (low ε at thermal IR) and emits little inside.

Where Stefan-Boltzmann shows up

• Stellar astrophysics. Effective temperatures from spectra times σT⁴ give surface fluxes; multiplied by 4πR² they give luminosities. The Hertzsprung-Russell diagram, the entire stellar-evolution framework, and exoplanet habitable-zone calculations all run on Stefan-Boltzmann conversions between T and L.
• Climate and Earth's energy budget. Earth radiates about 240 W/m² (averaged over the globe) at an effective temperature of 255 K. The greenhouse effect raises surface T by 33 K above this radiative equilibrium. Doubling atmospheric CO₂ adds about 4 W/m² of forcing, which by σT⁴ differential corresponds to ~1.2 K of warming before feedbacks.
• Infrared thermography. A thermal camera measures power emitted in 8–14 μm; assuming high-emissivity surfaces (ε ≈ 0.95 for skin, paint, rubber) it inverts σT⁴ to recover surface temperature. Used in building-envelope inspection, electrical-equipment diagnostics, fever screening at airports and chip-package failure analysis.
• Industrial furnace and kiln design. A 1500°C ceramic kiln radiates 5.67×10⁻⁸ × (1773)⁴ ≈ 560 kW/m². The wall losses and burner sizing follow directly from this. Pyrometric "optical" temperature measurement reads filament colour and inverts σT⁴ to get filament temperature within ±5 K up to 3000°C.
• Spacecraft thermal control. A satellite in deep space has only one heat-rejection mechanism: radiation. White paint with ε ≈ 0.85 and α_solar ≈ 0.20 is the work-horse coating: high emission of waste heat at IR, low absorption of sunlight. The James Webb telescope's tennis-court-sized sunshield exploits the same trade-off, dropping the cold-side temperature to 40 K.

The companion: Wien's displacement law

Stefan-Boltzmann tells you the total power radiated; Wien's displacement law tells you where in wavelength it peaks. The peak wavelength of a blackbody spectrum is:

λ_max × T = 2.898 × 10⁻³ m·K

Sun (5778 K): λ_max = 502 nm — green light, in the middle of the visible spectrum (and where human eyes are most sensitive, which is not a coincidence). Tungsten filament (3000 K): λ_max = 966 nm — near-infrared, with only 5 % of total power in the visible, which is why incandescent bulbs are inefficient. Human body (310 K): λ_max = 9.4 μm — thermal infrared, the band IR cameras image. Cosmic microwave background (2.7 K): λ_max = 1.07 mm — microwave wavelengths, exactly where Penzias and Wilson detected it in 1965.

The two laws together — σT⁴ for total power, λ_max T = b for peak wavelength — capture almost everything an engineer needs about thermal radiation without invoking the full Planck spectrum. For systems where wavelength selectivity matters (solar absorbers, IR detectors, atmospheric windows) the full Planck distribution is necessary; for total energy budgets, Stefan-Boltzmann alone usually suffices.

Variants and extensions

• Grey-body emission P = εσT⁴. The most common engineering generalisation. Emissivity ε < 1 is treated as wavelength-independent. Adequate for ε > 0.7 surfaces in the thermal IR; less so for selective surfaces like solar collectors.
• Two-surface radiative exchange. Net heat flow between two surfaces at T₁ and T₂ is q₁₂ = ε_eff σ (T₁⁴ − T₂⁴), where ε_eff combines both emissivities and view factor geometry. Drives the design of vacuum dewars and multilayer insulation.
• Wien's displacement law λ_max T = b. The peak-wavelength companion. Together with Stefan-Boltzmann it gives the two integrated properties of the Planck spectrum without solving the full integral.
• Selective spectral surfaces. Solar absorbers, telescope mirrors and radiative-cooling roofs deliberately tune ε(λ) to be high in one band and low in another — beating naive Stefan-Boltzmann arguments. Passive radiative coolers exploit the 8–13 μm atmospheric window to drop surface temperatures below ambient even in sunlight.
• Non-equilibrium thermal emission. When emitter and surroundings are not in thermal equilibrium (LEDs, fluorescent tubes, bioluminescence, hot plasmas), the Stefan-Boltzmann form breaks down — emission is not a single-temperature blackbody. Quantum-cascade lasers and IR sources designed with photonic crystals can radiate selectively with effective ε > 1 in narrow bands, exceeding what naive σT⁴ predicts.

Common pitfalls

• Using temperature in Celsius instead of kelvin. The fourth-power makes this mistake catastrophic: 50°C ≠ 323 K, but using 50 instead of 323 underpredicts radiated power by a factor of (50/323)⁴ ≈ 0.0006 — a thousandfold error. Always convert to kelvin first.
• Forgetting to include emissivity for real surfaces. A polished aluminium pot at 100°C radiates only 4 % of σT⁴, not 100 %. Engineering heat-loss calculations that ignore ε can be wrong by 25× for low-emissivity surfaces.
• Treating σT⁴ as the only heat-loss term. Radiation competes with convection and conduction. At 30°C an indoor surface loses heat about half by radiation and half by free convection. Heat-balance equations need all three pathways, summed coherently with their respective heat-transfer coefficients.
• Confusing irradiance with radiative flux. Stefan-Boltzmann gives the flux emitted by a surface into a hemisphere. The irradiance arriving on a receiving surface depends on geometry — view factors, distance, angle. Beware mistaking 6.3×10⁷ W/m² (Sun's surface flux) for 1361 W/m² (irradiance at Earth's distance); these differ by the inverse-square geometry factor.
• Applying the law to non-thermal radiation. Stefan-Boltzmann describes thermal-equilibrium blackbody emission. It does not describe lasers, LEDs, fluorescence, synchrotron radiation, or any non-thermal source. Each of those obeys its own emission physics; σT⁴ is irrelevant.