Planck’s law specifies how the intensity of blackbody radiation depends on frequency or wavelength and temperature. A blackbody is an ideal absorber that, in thermal equilibrium, emits a universal spectrum determined solely by its absolute temperature. Proposed by Max Planck in 1900, the law explains both the long-wavelength behavior described by classical physics and the rapid decrease of thermal radiation at high frequencies. Its derivation introduced an elementary energy scale proportional to frequency, a decisive step toward quantum mechanics.
The law is fundamental to thermodynamics, statistical mechanics, and the measurement of thermal radiation. It underlies radiation thermometry, the interpretation of stellar and planetary spectra, and the description of the cosmic microwave background. Its practical use requires careful distinctions among spectral radiance, energy density, and emitted power, as well as recognition that real materials generally emit less radiation than an ideal blackbody.
Mathematical statement and radiometric quantities
For radiation in vacuum, Planck’s law in frequency form is
Here is the spectral radiance, is frequency, is absolute temperature, is the Planck constant, is the Boltzmann constant, and is the speed of light in vacuum. Spectral radiance specifies power per projected emitting area, per solid angle, and per frequency interval. Its SI unit is . Thus, radiation leaving an area into solid angle , at angle to the surface normal, carries power
The wavelength form is
When wavelength is expressed in metres, has units . Values quoted per micrometre differ numerically by a factor of . The frequency and wavelength forms describe the same radiation but are different spectral densities: their numerical values cannot be interchanged simply by substituting . Conservation of power in corresponding intervals requires
For an isotropic equilibrium radiation field, the spectral energy density is
The spectral radiant exitance—the power emitted into the outward hemisphere per unit surface area—is instead . The factor results from integrating over a hemisphere. Confusing this hemispherical flux with radiance produces common factors-of- errors.
Under the International System of Units definitions effective since 2019, , , and are exact. Consequently, the ideal Planck spectrum contains no adjustable material constant: once temperature and the spectral coordinate are specified, its absolute radiance is fixed.
Experimental origins and the quantum hypothesis
The blackbody problem emerged from nineteenth-century efforts to relate heat, light, and electromagnetic theory. In 1859–1860, Gustav Kirchhoff established that, at thermal equilibrium, the ratio of emission to absorption is universal when frequency and temperature are fixed. This result separated the universal radiation spectrum from the detailed composition of the emitting material. An opaque cavity with a small aperture provided a practical approximation: radiation entering the hole undergoes repeated absorption and reflection, making escape unlikely, while radiation leaving the hole samples the cavity’s equilibrium field.
Wilhelm Wien developed a displacement relation in 1893 and proposed a spectral distribution in 1896 that worked well at short wavelengths. Improved measurements by Otto Lummer and Ernst Pringsheim, and especially long-wavelength experiments by Heinrich Rubens and Ferdinand Kurlbaum, revealed departures from Wien’s distribution. These experiments were associated with the precision-radiometry program of Germany’s Physikalisch-Technische Reichsanstalt. They supplied a demanding test because a successful formula had to account for both the already established short-wavelength behavior and the newly measured infrared behavior.
Planck presented an interpolation formula on 19 October 1900 that reproduced the observed spectrum. On 14 December 1900, he presented a statistical argument involving discrete energy elements of magnitude
His original reasoning concerned the distribution of energy among idealized material resonators exchanging radiation with the electromagnetic field. It did not yet amount to the modern assertion that freely propagating light consists of particles. The historical distinction matters: the mathematical step that introduced energy quantization preceded a full physical interpretation of the radiation field.
In 1905, Albert Einstein proposed that light itself could display localized energy quanta, applying the idea to phenomena including the photoelectric effect. Satyendra Nath Bose’s 1924 treatment later obtained the Planck distribution through a new counting of photon states, forming the basis of Bose–Einstein statistics. The modern derivation therefore differs conceptually from Planck’s original resonator argument, even though it yields the same spectrum.
Classical alternatives also clarified the problem. Lord Rayleigh derived a low-frequency expression in 1900, and James Jeans contributed to its subsequent development and correction. The resulting Rayleigh–Jeans law predicts unlimited total radiation energy when extended to arbitrarily high frequency. Paul Ehrenfest introduced the term ultraviolet catastrophe in 1911. This chronology distinguishes the later textbook contrast between classical and quantum theories from the actual experimental route by which Planck reached his law.
Derivation from electromagnetic modes and quantum statistics
A modern derivation begins with the normal modes of the electromagnetic field inside a large cavity. The cavity walls impose boundary conditions, restricting the allowed spatial patterns of the field. In the large-volume limit, the discrete mode spectrum can be treated as continuous. Counting the allowed wave vectors and the two independent transverse polarizations gives the number of modes per unit volume between and :
This density of states is a classical result obtainable from Maxwell’s equations. The departure from classical physics enters through the mean thermal energy assigned to each mode.
A mode of frequency behaves as a quantum harmonic oscillator. Ignoring its temperature-independent ground-state contribution for the moment, its allowed excitation energies are , where . At temperature , the Boltzmann distribution weights these states by , with . The mode’s partition function is the geometric series
Differentiation gives its mean thermal energy:
Multiplication by yields , and the isotropic-field relation then yields Planck’s law. The derivation separates two ingredients: the number of available electromagnetic modes and the quantum-statistical energy carried by each mode.
Equivalently, a mode contains a mean number of photons
Photons are bosons, so many may occupy the same mode. In ordinary blackbody equilibrium their number is not conserved: the surrounding matter can create and absorb photons. Their equilibrium chemical potential is therefore zero. A radiation field with a constrained photon number need not have the ordinary Planck distribution.
The oscillator’s full quantum energy spectrum includes zero-point energy, . The corresponding contribution is not part of the thermal spectrum in Planck’s law. It does not vanish at zero temperature and is not an additional outward thermal flux that a cold object emits. Distinguishing thermal occupation from vacuum fluctuations is essential when connecting blackbody radiation with quantum field theory.
Classical and high-frequency limits
The dimensionless ratio
controls the character of the spectrum. It compares the energy of a single photon with the thermal energy scale. When , the expansion gives the Rayleigh–Jeans law:
In this regime the mean energy per mode approaches , agreeing with classical equipartition. Quantum discreteness becomes relatively unimportant because the typical occupation number is large: .
The classical approximation nevertheless cannot be extended over the entire spectrum. Because the number of modes grows as , its predicted energy density grows without bound at high frequency. Integrating the Rayleigh–Jeans expression to infinite frequency produces an infinite result at any nonzero temperature. Planck’s denominator prevents this divergence by strongly suppressing the occupation of modes whose quantum energy greatly exceeds the thermal scale.
At , the subtraction of 1 in the denominator becomes negligible, giving Wien’s high-frequency approximation:
or
The exponential factor dominates the polynomial increase in the prefactor. High-frequency photons are therefore rare in thermal equilibrium unless the temperature is sufficiently high. This suppression explains why an object at room temperature emits strongly in the infrared but negligibly in visible light.
The approximation errors can be assessed directly. At fixed frequency, the ratio of the Rayleigh–Jeans prediction to the exact Planck radiance is ; it always exceeds 1 for positive . The ratio of the Wien approximation to the exact result is , so the Wien approximation slightly underestimates the radiance. These relationships are useful when deciding whether a simplified expression meets a specified accuracy.
The relevant regime depends on both wavelength and temperature, not on wavelength alone. At , radiation at has , far outside the Rayleigh–Jeans limit. Conversely, centimetre-wavelength radiation can lie in that limit even for a source only a few kelvin above absolute zero. In precision microwave work, photon energies small by everyday standards can still require a measurable quantum correction.
Integrated radiation and thermodynamic properties
Integrating Planck’s law over all frequencies produces the Stefan–Boltzmann law. For a blackbody surface, the total radiant exitance is
where
The substitution separates the temperature dependence from the spectral shape. The remaining dimensionless integral is
Josef Stefan inferred the fourth-power relation experimentally in 1879, and Ludwig Boltzmann derived it thermodynamically in 1884. Planck’s law subsequently supplied its full spectral basis and expressed the proportionality constant in terms of microscopic constants.
The total equilibrium radiation energy density is
Radiation also carries momentum. In an isotropic photon gas, averaging the directional momentum flux gives pressure . This is an equation of state for radiation, distinct from that of a nonrelativistic material gas. Together with the zero chemical potential, it gives the entropy density
At fixed volume, the radiation contribution to heat capacity is .
Integrating the photon occupation rather than photon energy gives the photon number density:
where . The mean photon energy consequently is
This is not the photon energy at either the frequency-spectrum peak or the wavelength-spectrum peak. A distribution’s mean and its mode are different quantities.
For a black surface exchanging radiation with a large black enclosure, both at uniform temperatures and with full mutual exposure, the net outward power per area is . The second term represents absorbed environmental radiation. A blackbody emits approximately , but an identical object inside an enclosure at the same temperature has zero net radiative heat loss. Planck’s law describes emission and equilibrium intensity; it does not imply that every emitting body is undergoing net cooling.
Spectral maxima, scaling, and color
The wavelength at which reaches its maximum follows Wien’s displacement law:
Differentiating the wavelength form of Planck’s law gives the equation
whose relevant positive root is . The corresponding maximum of , however, satisfies
with root . Therefore
These two peak locations do not obey . They are maxima of densities defined with respect to different spectral coordinates.
A graph using equal logarithmic intervals has another natural representation. Since , the radiance per logarithmic frequency interval is . Correspondingly, the radiance per logarithmic wavelength interval is , and these two quantities agree at matching frequency and wavelength. Their peak satisfies , with . Thus, a claim about “the peak of a blackbody spectrum” is incomplete unless the plotted spectral quantity is specified.
Temperature also produces exact scaling relationships. In frequency form,
where the same function applies at every temperature. In wavelength form,
Increasing temperature shifts the spectrum toward higher frequencies and raises its intensity. The integrated output grows as , but the height of the peak grows as for frequency density and as for wavelength density. These different exponents are consistent because the characteristic spectral interval also changes with temperature.
At , the wavelength-density maximum is near . For a blackbody at approximately , close to the effective temperature of the Sun, it is near . The latter does not mean that the Sun emits only green light or appears green. A thermal spectrum is broad, and perceived color depends on the combined stimulation of the eye’s receptors across the visible range.
Color temperature characterizes a spectrum by comparison with blackbody radiation. It can describe the visible appearance of a lamp even when the lamp is not in thermal equilibrium and does not follow Planck’s law over its full spectrum. A matching chromaticity therefore does not establish a matching physical temperature, total radiant output, or infrared emission.
Real surfaces and radiative transfer
A real surface usually emits less radiation than a blackbody at the same temperature. Its spectral directional emissivity is defined through
For an ordinary passive surface in the far field, emissivity is between zero and one. It can depend on frequency, direction, polarization, temperature, surface roughness, and chemical condition. A gray-body model assumes that emissivity is independent of frequency over the range under consideration. That assumption is a useful approximation, not a general property of matter.
Kirchhoff’s law of thermal radiation equates emissivity and absorptivity for corresponding optical modes under the usual equilibrium and reciprocity conditions. A surface that absorbs poorly at a particular wavelength also emits poorly there. For an opaque surface, transmission vanishes, and absorptivity is one minus reflectivity when the quantities are consistently defined. Polished metals therefore commonly have low thermal emissivity and reflect a substantial part of the surrounding infrared field. Oxidation or a coating can change this behavior markedly.
The radiance observed from a surface is not necessarily just its own thermal emission. Reflected environmental radiation may contribute, and a semitransparent object may transmit radiation originating behind it. Even a low-emissivity surface can consequently appear bright to an infrared instrument if it reflects a hot source. Surface temperature cannot generally be inferred from measured radiance without a model of these contributions.
For an absorbing and emitting medium, the connection is expressed by radiative transfer. Neglecting scattering, the intensity along a ray obeys
where is the absorption coefficient and is the emission coefficient. In local thermodynamic equilibrium,
For an isothermal layer with optical depth and incident intensity , the solution is
An optically thick layer approaches the Planck intensity, whereas an optically thin layer can be much fainter.
This distinction is central to gases. A hot gas does not automatically emit a smooth blackbody spectrum: molecular and atomic transitions can produce strong spectral structure, and an optically thin gas need not absorb enough radiation to establish a Planckian field. In local thermodynamic equilibrium, Planck’s law supplies the local emission-to-absorption ratio, even when the radiation propagating through the entire system is not itself in equilibrium. Temperature gradients, selective absorption, and scattering then determine the spectrum that escapes.
Radiation thermometry and calibration
Radiation thermometry determines temperature from emitted electromagnetic radiation rather than by direct thermal contact. Optical pyrometers historically compared an incandescent source with a calibrated lamp; modern infrared instruments measure radiance in selected spectral bands. Planck’s law provides the theoretical relation between temperature and ideal blackbody radiance, while source emissivity, the surrounding radiation field, and instrument response determine how the ideal relation is applied.
For a spectrally narrow measurement, a brightness temperature is the blackbody temperature that would produce the observed radiance at the specified frequency:
This is an equivalent radiometric temperature, not necessarily the material’s thermodynamic temperature. An emitting surface with emissivity below one, without compensating reflected radiation, has a brightness temperature below its actual temperature. In radio astronomy, “brightness temperature” often instead denotes the Rayleigh–Jeans equivalent ; the convention must therefore be stated.
Real detectors respond over a finite interval. A simplified instrument signal has the form
where includes spectral responsivity and, as appropriate, optical throughput. Replacing this integral with the radiance at a nominal central wavelength can introduce temperature-dependent errors because the spectrum changes shape with temperature. Photon-counting detectors additionally require the conversion from radiant energy to photon number, involving the factor .
High-quality calibration uses sources that approximate blackbody cavities. A small aperture, absorbing walls, and multiple internal reflections can produce an effective aperture emissivity close to unity even when the wall material itself is imperfect. Temperature uniformity remains crucial: radiation emerging from a cavity with gradients is a weighted combination of contributions at different temperatures, not exactly the spectrum of a single temperature. Aperture geometry, diffraction, stray light, detector nonlinearity, and the measured area also enter the measurement uncertainty.
The logarithmic sensitivity of monochromatic Planck radiance to temperature is
It approaches 1 in the Rayleigh–Jeans regime and approximately in the Wien regime. Short-wavelength radiance can therefore be highly sensitive to temperature, although the signal may be weak for cool objects. This sensitivity is useful in metrology, but it does not remove the need to characterize emissivity and the optical path.
Astronomy, cosmology, and planetary radiation
Planck’s law is a basic interpretive tool in astronomy because many astronomical objects contain regions that are approximately opaque and in local thermodynamic equilibrium. A stellar photosphere can produce a continuum resembling a blackbody, with spectral lines and continuum departures imposed by its composition, opacity, and temperature structure. A star’s effective temperature is defined through its luminosity and radius:
This definition does not assert that the complete emergent spectrum is exactly Planckian. Spectral color temperatures, brightness temperatures, and effective temperature can differ for the same star.
The cosmic microwave background provides an especially accurate realization of an astronomical blackbody spectrum. Measurements by the Far Infrared Absolute Spectrophotometer aboard the Cosmic Background Explorer satellite established its nearly Planckian form. Its present temperature is approximately . At this temperature, peaks near , whereas peaks near . The difference is the expected consequence of the spectral-coordinate transformation, not a disagreement between measurements.
Cosmic expansion preserves the shape of a Planck spectrum during free propagation while reducing its temperature. Photon frequencies decrease through cosmological redshift, and photon number density decreases as space expands. The invariant phase-space occupation implies
Processes in the early universe that exchanged energy or created and absorbed photons helped establish the spectrum. Later energy release can instead produce distortions, making the accuracy of the measured Planck form a probe of cosmic thermal history. A spectrum with a chemical-potential distortion, for example, is not described by an ordinary blackbody at merely a different temperature.
Thermal emission from interstellar dust is often represented by a modified blackbody, with an optically thin approximation
A frequency-dependent optical depth, frequently modeled over a limited range by a power law, changes the spectral slope. Multiple dust temperatures introduce further departures from a single Planck curve. The fitted temperature can consequently depend on the spectral interval and on the assumed dust emissivity.
Planetary radiation involves the same distinction between local emission and the outgoing spectrum. Earth’s surface and atmosphere emit in the infrared, but absorption and emission by gases and clouds cause different wavelengths to escape from different heights and temperatures. Planck’s law supplies the local thermal source function, while radiative transfer determines the outgoing spectrum and its contribution to the Earth’s energy budget. A single effective radiating temperature expresses the total outgoing power; it does not describe every wavelength or every atmospheric layer.
Conditions of validity and extensions
The standard Planck formula describes the propagating electromagnetic modes of an equilibrium radiation field in vacuum, or the ideal blackbody radiance emitted into vacuum. Its universality depends on the radiation attaining thermal equilibrium through suitable interactions with matter. A source being hot is not enough. Radiation produced by lasers, electrical discharges, fluorescence, or accelerated charged particles can have spectral and statistical properties unrelated to a Planck distribution, even when parts of the source have a well-defined material temperature.
Local thermodynamic equilibrium is a less restrictive and widely useful condition. Matter in a small region can have approximately thermal level populations at a local temperature, giving a Planckian thermal source function, while the radiation traveling through that region remains non-Planckian. When excitation, ionization, or emission is controlled by processes that do not maintain thermal populations, a non-equilibrium treatment is necessary. Detailed atomic or molecular transition rates then replace a simple assumption that the source function equals .
The optical environment also matters. The factor assumes the mode density of a large vacuum volume. Inside a dielectric, wave propagation and the density of available states depend on material properties, including refractive index and dispersion. In a nondispersive, transparent medium, familiar refractive-index factors can be derived, but a single vacuum expression cannot be applied indiscriminately inside absorbing or strongly dispersive matter. Transfer across an interface must also account for refraction, reflection, and the relevant propagation directions.
At separations much smaller than the dominant thermal wavelength, near-field heat transfer can involve evanescent electromagnetic fields. Coupling through these fields, sometimes enhanced by surface resonances, can produce heat fluxes exceeding the conventional far-field blackbody exchange between parallel surfaces. This does not contradict Planck’s law: the far-field bound counts propagating vacuum channels, whereas near-field transfer includes additional channels that do not propagate to distant observers. Likewise, the absorption or emission cross-section of a small resonant particle need not equal its geometrical area, so apparent “super-blackbody” behavior requires a precise definition of area and geometry.
Relativistic observations require another qualification. Spectral radiance is frame-dependent, but is invariant under Lorentz transformations. An observer moving relative to an isotropic blackbody field sees direction-dependent Doppler shifts and an associated direction-dependent Planck temperature. A gravitational frequency shift similarly preserves the Planck shape along a ray while changing the inferred temperature. In a static gravitational field, global thermal equilibrium requires the locally measured temperature to vary with the gravitational time-dilation factor, as described by the Tolman relation. These extensions preserve the occupation-statistical structure of Planck’s law while specifying the frame, geometry, and equilibrium conditions under which temperature is assigned.