Radiative transfer

Radiative transfer describes emission, absorption, and scattering of radiation through matter.

Multiplying an absorber abundance by $a$ with fixed cross sections and physical temperature profile multiplies optical depth by $a$. The Planck function as a function of the new coordinate must therefore be composed with division by $a$. A term proportional to ${\tau }^n$ acquires coefficient $a^{-n}$. Holding its old optical-depth coefficient fixed would instead change the physical temperature profile. This is the immediate radiative comparison before thermal adjustment.

A scattering phase function describes the angular distribution of single-scattered power. One common convention is $\int Pd\Omega =4\pi$, so that $P/(4\pi )$ is a probability density over outgoing solid angle. For unpolarized Rayleigh scattering, $P=3(1+{\cos }^2\Theta )/4$. It is distinct from the oscillatory-integral meaning of phase function.

The scattering asymmetry parameter is the mean scattering-angle cosine, $g=(4\pi {)}^{-1}\int P\cos \Theta d\Omega$ in the $4\pi$ normalization. Forward/backward symmetry gives $g=0$; positive $g$ indicates a net forward preference. In a two-stream transport approximation, the reduced scattering optical depth is $(1-g)\tau$.

The derivative of outgoing spectral radiance with respect to a pressure-resolved absorber abundance defines a kernel once the vertical measure is fixed. In an optically thin column with $k\propto p$ and $B/B_0=p/p_0$, $K_p=[k_0{B}_0/(p_{0}gM)]p(p/p_0-1)$. It is negative inside the atmosphere and zero at both endpoints. Its largest magnitude lies at $p_0/2$, whereas the logarithmic-pressure kernel $p{K}_p$ peaks in magnitude at $2p_0/3$. A uniformly perturbed abundance instead has one integrated derivative, not a pressure location.

A two-stream approximation replaces the angular distribution of radiation by upward and downward representative intensities or hemispheric fluxes, with a specified angular closure. Angular path factors can be absorbed into an effective optical depth, producing exponential attenuation in that coordinate. Different closures change numerical emission depths and diffusivity factors. Fluxes and directional radiances must remain distinguished, and a prescribed effective optical depth need not equal literal vertical absorption optical depth.

If optical depth is $\tau (z)={\tau }_s{e}^{-z/H}$, multiplying opacity by $a$ raises any fixed-optical-depth surface by $H\ln a$. In a constant lapse-rate layer this changes its temperature by $-\Gamma H\ln a$.

An absorption equivalent width is $W_{\lambda }=\int [1-I_{\lambda }/I_{\lambda }^0]d\lambda$, the wavelength width of a completely black interval that removes the same continuum-normalized area. For a narrow spectral line profile, small optical depth and negligible stimulated emission, the energy-density convention for Einstein coefficients gives $W_{\lambda }=h{B}_{12}{N}_1/{\nu }_{12}$. This follows by replacing $1-e^{-{\tau }_{\nu }}$ by ${\tau }_{\nu }$ and using $|d\lambda /d\nu |=c/{\nu }^2$. At redshift $z$, the observed and rest-frame values satisfy $W_{\mathrm{obs}}=(1+z)W_{\mathrm{rest}}$.

A curve of growth describes the equivalent width $E(m)=\int (1-e^{-m{\kappa }_{\nu }})d\nu$ as a function of absorber mass path or column density. Its weak-line limit is $E\simeq mS$. Saturation makes growth sublinear: a pure Gaussian core gives $E\sim 2{\Delta }_D\sqrt{\log {\tau }_0}$, while saturated Lorentzian line shape wings give $E\sim 2\sqrt{mS{\Delta }_L}$. The spectral coordinate and coefficient convention must be consistent.

The emission coefficient $j_{\nu }$ measures emitted energy per unit volume, time, solid angle and frequency. Its SI units are $\mathrm{W}{\mathrm{m}}^{-3}\mathrm{s}{\mathrm{r}}^{-1}\mathrm{H}{\mathrm{z}}^{-1}$. Along a ray it contributes $j_{\nu }ds$ to specific intensity. For an isotropic spontaneous emission transition with a normalized spectral line profile, $j_{\nu }=h{\nu }_{21}{n}_2{A}_{21}\phi (\nu )/(4\pi )$; integration over solid angle removes the $4\pi$ denominator.

If fraction $a$ of scattered energy is exactly forward and the rest isotropic, a single-escape slab has upward fraction $(1-a)(1-t)/2$ and downward fraction $t+[a+(1-a)/2](1-t)$. They sum to one for conservative scattering; this approximation need not be accurate for a thick cloud.

If scattering is elastic and there is no absorption or thermal emission in the band, the integrated outgoing energy equals incoming energy. Consequently net radiative-flux convergence and matter heating vanish even when the direct beam is almost extinguished.

Taking the first angular moment of the radiative-transfer equation gives $d{F}_{\nu }/dz=4\pi {\alpha }_{\nu }(S_{\nu }-J_{\nu })$. Radiative equilibrium sets the frequency integral of the right-hand side to zero, so the bolometric flux is constant with depth.

The Planck function is the equilibrium spectral specific intensity of a blackbody, given by Planck's law. Per unit wavelength it isIt can also be written per unit frequency, with $B_{\nu }d\nu =-B_{\lambda }d\lambda$ under the corresponding change of variable.

The radiation constant relates equilibrium radiation energy density to temperature by $u=a_{\mathrm{rad}}{T}^4$. It satisfies $a_{\mathrm{rad}}=4\sigma /c$ in terms of the Stefan-Boltzmann constant.

The Rayleigh-Jeans law is the long-wavelength, low-frequency limit of Planck law. When $hc/(\lambda k_{B}T)\ll 1$,

The ultraviolet catastrophe is the divergent total thermal radiation energy predicted by classical equipartition theorem applied to the electromagnetic mode density in a box. Each mode has mean energy $k_{B}T$, while the number of modes per volume per frequency is $8\pi {\nu }^2/c^3$. The Rayleigh-Jeans law therefore gives $u_{\nu }=8\pi k_{B}T{\nu }^2/c^3$, whose integral over all positive frequencies diverges at every $T>0$.

Planck's law replaces the classical mode energy by $h\nu /(e^{h\nu /(k_{B}T)}-1)$. It agrees with equipartition when $h\nu \ll k_{B}T$ but suppresses high-frequency excitations exponentially. The resulting spectral energy density is integrable and yields the finite Stefan–Boltzmann law. The failure concerns the classical equilibrium distribution of energy among modes, not the existence of electromagnetic wave propagation itself.

The Wien approximation is the short-wavelength limit of the Planck law:

Wien's displacement law says that the wavelength maximizing the Planck function per unit wavelength, $B_{\lambda }(T)$, moves inversely with absolute temperature:Here $b$ is the Wien displacement constant. Heating an ideal blackbody therefore shifts this peak toward shorter wavelengths, while also increasing its emission at every fixed wavelength. The peak identifies where equal narrow wavelength intervals carry the most power; it is not a single emitted wavelength. Planck's law supplies the full curve, and the Stefan–Boltzmann law supplies its integral.

Specify the spectral coordinate before finding a peak. The frequency form of Wien's displacement law maximizes $B_{\nu }$, which is spectral radiance per unit frequency. Its maximum is not $c/{\lambda }_{\max }$. The frequency and wavelength spectral radiance conversion changes the density as well as the horizontal coordinate; the existing frequency and wavelength peaks of Planck's law give a complementary numerical comparison.

In his 11 December 1911 Nobel lecture, Wilhelm Wien explained the displacement principle through a thought experiment: compress black-body radiation between reflecting moving walls. Work against radiation pressure raises the energy density, while reflection changes each frequency by the Doppler effect. The Second law of thermodynamics connects the resulting spectrum to one at a higher temperature. This determines how the spectrum scales between temperatures, without determining its complete shape. The Wien approximation is a separate, restricted spectral formula.^[ref]

At fixed $T>0$, substitute $x=hc/(\lambda k_{B}T)>0$ into the Planck function:Here $h$ is the Planck constant, $k_B$ the Boltzmann constant, and $c$ the speed of light. Since $dx/d\lambda =-x/\lambda \ne 0$, an interior maximum in wavelength is also a stationary point in $x$. A logarithmic derivative givesThe required positive root is $x_{\lambda }=4.9651142317\dots$. The rearranged equation also has $x=0$, but that endpoint is not a peak. For $g(x)=5(1-e^{-x})-x$, $g^{\prime }(x)=5e^{-x}-1$ changes sign once, at $x=\log 5$. Thus $g$ first increases and then decreases to negative values, crossing zero exactly once for $x>0$. Since $x^5/(e^x-1)$ tends to zero at both endpoints, this stationary point is the unique global maximum. Solving the definition of $x$ for wavelength givesThe inverse-temperature dependence follows before any numerical root-finding. More generally $B_{\lambda }(T)=T^{5}F(\lambda T)$ for one fixed function $F$, so spectra at all temperatures have the same rescaled shape. This is stronger than a relation between two peak positions. Replacing the denominator by $e^x$ in the Wien approximation would instead give $x=5$ and a peak wavelength about $0.70\%$ too short.

For Planck's law per unit frequency, setting $x=h\nu /(k_{B}T)$ leaves a positive constant times $x^3/(e^x-1)$. Its logarithmic derivative givesThe same uniqueness argument as in the Derivation of Wien's displacement law, with $5$ replaced by $3$, selects the positive root. The maxima differ because corresponding positive spectral intervals obeyThe nonconstant Jacobian determinant $c/{\lambda }^2$ changes density heights. Consequently $c/{\nu }_{\max }=(x_{\lambda }/x_{\nu }){\lambda }_{\max }\simeq 1.75978{\lambda }_{\max }$. A peak is a property of a specified density, rather than an invariant point of the spectrum. For equal logarithmic intervals, the corresponding densities satisfy $\lambda B_{\lambda }=\nu B_{\nu }$ and have matching converted peak positions; this is a third convention, distinct from either linear-interval density.

The Wien displacement constant is the constant of proportionality $b$ in ${\lambda }_{\max }T=b$. From the exact International System of Units values of the Planck constant, speed of light and Boltzmann constant,The units follow from $hc/k_B$: $(\mathrm{J}\mathrm{s})(\mathrm{m}{\mathrm{s}}^{-1})/(\mathrm{J}{\mathrm{K}}^{-1})=\mathrm{m}\mathrm{K}$. The defining constants are exact; the displayed decimal is rounded. ^[ref]

Consider a hypothetical star whose emitted specific intensity is exactly that of a uniform $5000\mathrm{K}$ blackbody. Assume transparent propagation with negligible Doppler shift and interstellar extinction, so geometric dilution multiplies every spectral interval by the same factor. The stipulated temperature givesUsing the converted frequency peak as though it maximized $B_{\lambda }$ would return $b/(c/{\nu }_{\max })\simeq 2841.26\mathrm{K}$, a substantial error caused solely by mixing spectral conventions.

Now stipulate a photodetector with constant quantum efficiency $\eta$ over the relevant wavelengths, collecting equal narrow wavelength bins. For fixed accepted projected area times solid angle $G$, with dilution included in $G$, its expected photon count rate density isThis follows by dividing the received power by the photon energy $hc/\lambda$. It peaks where $x=4(1-e^{-x})$, at $x=3.9206903949\dots$, giving $0.73394057\mu \mathrm{m}$ for the same model star. Constant quantum efficiency is not constant energy spectral responsivity: the latter includes the wavelength-dependent photon energy conversion. For example, $G=10^{-12}{\mathrm{m}}^2\mathrm{sr}$, $\eta =0.5$ and a $1\mathrm{nm}$ bin centered at $0.600\mu \mathrm{m}$ give approximately $1.93\times 10^{10}$ detected photons per second in the narrow-bin approximation. These are calculated model values, not observations of a named star.

A more general linear detector channel has outputwhere $S_{i,0}$ is the background reading and $R_i$ includes spectral selection, optical transmission and detector spectral responsivity. A raw channel maximum need not locate either Planck function maximum. Recovering a temperature requires an appropriate calibrated response and an emission model. For a real stellar atmosphere, opacity and radiative transfer can reshape the spectrum; interstellar extinction can reshape it again. A single peak does not establish a unique physical temperature without those assumptions.

William Weber Coblentz's primary report describes an electrically heated porcelain cavity wound with platinum ribbon. Mirrors imaged its central radiating diaphragm onto a $0.6\mathrm{mm}$ optical slit; a fluorite optical prism dispersed the beam onto a vacuum bolometer. Galvanometer deflections read the detector response. A resistance standard checked electrical sensitivity, while repeated readings checked stability. Sodium lines fixed the angular zero; a refractive index curve supplied wavelengths and slit-bandwidth corrections. Thermocouples were standardized at freezing points of zinc, antimony and copper. Drying the optical path reduced atmospheric absorption; prism scattering and temperature calibration remained limitations. These were Coblentz's measurements. His extraction of radiation constants assumed the Planck law, so it was not a model-independent proof of the spectral formula.^[ref]

The inference to test is whether corrected spectra taken at independently determined temperatures have their $B_{\lambda }$ maxima at a constant product ${\lambda }_{\max }T$. Comparing raw galvanometer deflections at different optical prism settings would also compare different wavelength widths and optical losses. The response integral in blackbody peak and detector response explains why those corrections belong to the physical inference rather than merely to presentation.

A blackbody absorbs all incident electromagnetic radiation. In thermal equilibrium it emits the universal Planck law spectrum determined solely by its temperature.

The fraction of incident radiant power absorbed under stated spectral and angular conditions. At thermal equilibrium, Kirchhoff's law of thermal radiation equates matching directional spectral absorptivity and emissivity. A blackbody has both equal to one.

An isotropic blackbody surface has outward flux $F=\pi \int B_{\nu }d\nu =\sigma T^4$. A sphere of radius $R$ therefore has luminosity

Thermal bremsstrahlung is free-free electromagnetic radiation from the acceleration of charged particles during Coulomb encounters in a thermal plasma. Its frequency-integrated emissivity in a fully ionized, nonrelativistic plasma scales approximately as $n_e{n}_i{T}^{1/2}$.

For a uniform ionized hydrogen column with known temperature and fixed projected angular area, free-free brightness and optical depth depend on $\mathrm{EM}=n_e^{2}L$. The displayed rescaling leaves the whole ideal spectrum unchanged. density and path length cannot be separated without additional information. A known geometry/distance, or a distinct column diagnostic proportional to $n_{e}L$, can break the degeneracy; gas clumping factor and temperature structure complicate that inference.

The fraction $f_V$ of a geometric volume occupied by a specified emitting gas phase is its volume filling factor. For approximately uniform density within that phase, a density-squared luminosity scales as $n_e^2{f}_{V}V$. Without knowledge of $f_V$, emission gives an rms-density combination rather than a unique mean gas density. This filling factor is distinct from the fraction of a telescope beam occupied on the sky.

For a hydrogen-dominated fully ionized plasma, the line-of-sight emission measure is $\mathrm{EM}=\int n_e^{2}dl$. Its units distinguish it from a volume-integrated emission measure $\int n_e^{2}dV$. Thermal free-free optical depth is proportional to $\mathrm{EM}{T}^{-3/2}{\nu }^{-2}{\bar{g}}_{\mathrm{ff}}$ in the radio limit. Inferring density also requires depth, ionization and clumping assumptions.^[ref]

The emission coefficient per volume, frequency and solid angle of a Maxwellian plasma has $j_{\nu }\propto n_e{\sum }_i{Z}_i^2{n}_i{T}^{-1/2}{e}^{-h\nu /(k_{B}T)}{\bar{g}}_{\mathrm{ff}}$. Define the angular convention before supplying a numerical coefficient. The high-frequency cutoff measures temperature, while the normalization measures a density-squared emission integral. The radio spectrum is almost flat when optically thin and rises as ${\nu }^2$ when optically thick at fixed projected area.

For fully ionized hydrogen and helium with hydrogen mass fraction $x_H$, $n_e=(1+x_H)\rho /(2m_p)$ and ${\sum }_i{Z}_i^2{n}_i=\rho /m_p$. With the same unit Gaunt factor for both species, thermal bremsstrahlung at fixed electron number density is multiplied by $2/(1+x_H)$ relative to pure ionized hydrogen. The factor is one for pure hydrogen and two for pure helium.

The Gaunt factor is the dimensionless quantum correction to a classical free-free radiation calculation. Its weak temperature and frequency dependence modifies the approximate thermal bremsstrahlung scaling; a frequency-averaged value appears in the bolometric cooling coefficient.

Optical depth satisfies $d{\tau }_{\nu }={\alpha }_{\nu }ds$ along a path and measures attenuation through $I_{\nu }\propto e^{-{\tau }_{\nu }}$ in the absence of emission.

An optically thick medium has optical depth $\tau \gg 1$ along the relevant path. Photons undergo many interactions before escaping. Radiative diffusion applies under suitable near-isotropy conditions, but a large scattering depth alone does not guarantee thermalization when true absorption is weak.

An optically thin medium has $\tau \ll 1$, so attenuation and emission may be expanded to first order in optical depth.

For isotropic emission without significant absorption, integrating the spontaneous photon-production rate along the sightline gives the displayed energy intensity per solid angle. Photon intensity omits the factor $h\nu$. Summing unresolved transitions is necessary before comparing with an observed blended spectral feature.

The absorption coefficient is the fractional reduction of beam intensity per unit path length: $d{I}_{\nu }/ds=-{\alpha }_{\nu }{I}_{\nu }$ when emission and scattering are absent.

A collision or transient molecular pair can possess a dipole even when the isolated molecules do not. The resulting opacity can be important in a dense molecular hydrogen atmosphere. Pair absorption per unit length scales approximately as the product of the two partner number densities, so its mass absorption coefficient is generally density-dependent. It supplies continuum absorption between ordinary spectral lines.

For dust-grain number density $n_d$ and absorption cross-section ${\sigma }_d$, the dust optical depth along a path is ${\tau }_d=\int n_d{\sigma }_{d}ds$.

The mass absorption coefficient is absorbing cross-section per unit mass. For identical spherical dust grains of radius $a$, mass $m_d$, and absorption efficiency $Q_{\nu }$, it is ${\kappa }_{\nu }=\pi a^2{Q}_{\nu }/m_d$.

For isothermal optically thin dust at distance $D$, its measured spectral flux density determines its mass through $M_d=F_{\nu }{D}^2/[{\kappa }_{\nu }{B}_{\nu }(T)]$. Temperature and opacity uncertainty commonly dominate the inferred mass uncertainty.

In plane-parallel optical-depth coordinates without scattering,with source function $S_{\nu }$.

The absorption-and-emission transfer equation without scattering, often expressed along a ray as $d{I}_{\lambda }/ds=-{\alpha }_{\lambda }{I}_{\lambda }+{\alpha }_{\lambda }{B}_{\lambda }(T)$ in local thermal equilibrium. Its formal solution weights thermal emission by intervening transmission.

The radiative transfer source function is emission per unit extinction. If $j_{\nu }$ is emission per unit path length and $\rho {\kappa }_{\nu }$ is extinction per unit path length, $S_{\nu }=j_{\nu }/(\rho {\kappa }_{\nu })$ and the transfer equation is $\mu {\partial }_{{\tau }_{\nu }}{I}_{\nu }=I_{\nu }-S_{\nu }$. Under thermal absorption in local thermodynamic equilibrium it is the Planck function; for conservative coherent isotropic scattering it is the mean intensity.

In an optically thick medium whose radiation field is nearly isotropic and thermal, the first angular correction to the Planck function produces the diffusive fluxwhere ${\kappa }_R$ is the Rosseland mean opacity.

The Rosseland mean opacity is the harmonic frequency average weighted by $\partial B_{\nu }/\partial T$. It controls radiative diffusion because flux preferentially travels through low-opacity frequencies.

It is a frequency-averaged opacity appropriate to the radiative-diffusion limit, not the general opacity concept itself.

In local thermodynamic equilibrium, matter populations are thermal at the local temperature and the source function is the Planck function, $S_{\nu }=B_{\nu }(T)$.

For an outward ray through a nonscattering atmosphere,

Along a radial ray through a layer with constant source function, the incident specific intensity is attenuated by $e^{-\tau }$ and the layer adds $S(1-e^{-\tau })$. At small optical depth, $I_{\mathrm{out}}=I_{\mathrm{in}}+(S-I_{\mathrm{in}})\tau +O({\tau }^2)$. Whether the source exceeds the incident specific intensity determines the first-order emission or absorption contrast.

For a semi-infinite plane-parallel atmosphere with polynomial source function $S(\tau )=\sum a_j{\tau }^j$, the outward specific intensity is ${\int }_0^{\infty }S(\tau )e^{-\tau /\mu }d\tau /\mu$. Substitution $u=\tau /\mu$ gives coefficients $A_j=j!a_j$ in the surface polynomial in $\mu$. The deep boundary term must vanish. A local truncated Taylor series instead gives this contribution plus a remainder, so its surface polynomial is not automatically exact.

For a thermalized radiative transfer source function $S_{\lambda }=B_{\lambda }(T)$ constant along a ray, the specific intensity after optical-depth interval $\Delta \tau$ is $I_{\lambda }=B_{\lambda }+(I_{\mathrm{in}}-B_{\lambda })e^{-\Delta \tau /\mu }$ for $\mu >0$. As $\Delta \tau \to \infty$, it tends to the Planck function, independently of a bounded boundary input. A temperature gradient is unnecessary for blackbody emission. In a scattering medium, local thermodynamic equilibrium alone does not guarantee $S_{\lambda }=B_{\lambda }$; thermalization must be justified.

A non-scattering isothermal atmospheric layer in local thermodynamic equilibrium, of temperature $T_a$ and slant optical depth $t_{\lambda }$, lies above a blackbody background at $T_b$. The specific intensity isThus an increase in opacity shifts the specific intensity toward the layer's Planck function. If a line is more opaque than its continuum, it appears in absorption for $T_a<T_b$, in emission for $T_a>T_b$, and has no contrast for $T_a=T_b$. This is a local explanation, not a claim that all features diagnose one uniform layer.

When the source function varies approximately linearly with optical depth, the emergent intensity obeys $I_{\nu }(0,\mu )\simeq S_{\nu }({\tau }_{\nu }=\mu )$.

The relation is an approximation within radiative transfer, not the entire transfer theory.

Coherent isotropic scattering preserves photon frequency and redistributes direction uniformly. Its source function is the mean intensity $J_{\nu }$, so conservative scattering removes light from a pencil beam while preserving luminosity integrated over all escaping directions.

The single-scattering albedo is the fraction of total extinction due to scattering, ${\omega }_{\nu }={\alpha }_{\nu ,\mathrm{sca}}/{\alpha }_{\nu ,\mathrm{ext}}$. Its complement is the absorptive fraction.

Kirchhoff's law of thermal radiation states that a body in thermal equilibrium has equal absorptivity and emissivity at each frequency and direction. A layer with infrared absorptivity $ϵ$ therefore emits $ϵB_{\nu }(T)$.

For plane-parallel intensity, the standard moments areand the radiative flux is $F_{\nu }=4\pi H_{\nu }$.

The angular multipole of order two in a local radiation intensity distribution. A monopole is isotropic and a dipole has one directional axis; a quadrupole has a second-order angular pattern. Thomson scattering of an incident radiation quadrupole can generate linear polarization, which is central to Cosmic microwave background polarization.

The mean intensity is the angular average $J_{\nu }=(4\pi {)}^{-1}\int I_{\nu }d\Omega$ of the specific intensity.

Radiative flux is the first angular moment $F_{\nu }=\int I_{\nu }\cos \theta d\Omega$ of specific intensity and measures net radiant power crossing unit area. Integrating over frequency gives bolometric radiative flux. The receiving surface's orientation matters through $\cos \theta$; observed radiative flux from a compact distant source conventionally refers to an area perpendicular to its line of sight.

Outgoing longwave radiation is the outward thermal radiative flux at the top of the atmosphere, integrated over the specified thermal spectrum and outward directions. It differs from a directional spectral radiance.

For a steady point-like astronomical source with isotropy of emission, in transparent space with negligible relativistic or cosmological effects, conservation of energy spreads its luminosity over a sphere of surface area $4\pi d^2$. Its bolometric observed radiative flux isThis inverse-square law reflects geometric dilution: the same emitted power crosses increasingly large spheres. It does not require light to lose energy on the journey. For a fixed luminosity, doubling Euclidean distance quarters the observed radiative flux. ^[ref].

With $L_0=10^{26}\mathrm{W}$ and $d_0=20$ parsecs, the result is $F_0\simeq 2.09\times 10^{-11}\mathrm{W}{\mathrm{m}}^{-2}$. At $40$ parsecs, the same source gives $F_0/4\simeq 5.22\times 10^{-12}\mathrm{W}{\mathrm{m}}^{-2}$. An ideal collecting area of $1{\mathrm{m}}^2$ at $20$ parsecs intercepts approximately $2.09\times 10^{-11}\mathrm{W}$, before detector losses.

The distribution of received radiative flux or emitted luminosity across frequency or photon energy. Plotting $\nu F_{\nu }$ against $\log \nu$ makes the area equal to the integrated radiative flux, since $F_{\nu }d\nu =\nu F_{\nu }d\log \nu$. A convention such as $F_{\nu }$, $\nu F_{\nu }$ or $F_{\lambda }$ must be specified before comparing spectral slopes.

One observed radiative flux cannot determine both luminosity and Euclidean distance: for every $k>0$, replacing $(L,d)$ by $(k^{2}L,kd)$ leaves $F$ unchanged. In particular, $4\times 10^{26}\mathrm{W}$ at $40$ parsecs produces the same observed radiative flux as $10^{26}\mathrm{W}$ at $20$ parsecs. This is a concrete failure of identifiability in an inverse problem.

An independent stellar parallax fixes the distance and breaks this degeneracy; alternatively, a calibrated standard candle supplies the luminosity. Interstellar extinction can further reduce the observed radiative flux, and a photometric passband measures only a selected spectral range. At cosmological distances, use luminosity distance in the bolometric relation. Each inference must specify which radiative flux was measured and what propagation assumptions connect it to luminosity.

For a uniformly emitting spherical body of radius $R$ at Euclidean distance $d\gg R$, integration of its specific intensity over the apparent solid angle gives $f_{\lambda }=\pi I_{\lambda }(R/d{)}^2$. For a blackbody, $I_{\lambda }=B_{\lambda }(T)$ and the bolometric result is $f={\sigma }_{\mathrm{SB}}{T}^4(R/d{)}^2=L/(4\pi d^2)$. A grey emissivity factor multiplies this expression; isothermality alone does not imply blackbody emission.

Spectral flux density is radiant power received per unit area and per unit wavelength or frequency. The two conventions satisfy $F_{\lambda }d\lambda =-F_{\nu }d\nu$ under $\nu =c/\lambda$, hence $F_{\lambda }=(c/{\lambda }^2)F_{\nu }$ when the arguments refer to the same photons. Integrating over a photometric passband gives its measured flux.

The Eddington closure approximation sets the second angular moment to one third of the zeroth, so the radiation pressure and mean intensity satisfy $c{P}_r=4\pi J/3$.

For the affine angular ansatz $I=A+C\cos \theta$, the radiative flux is $F=4\pi C/3$. Setting the inward hemispheric flux to zero gives $A(0)=2C/3$, but then the formal intensity at $\cos \theta =-1$ is negative when $F>0$. Exact no-incoming radiation and nonnegative intensity cannot both be represented by a nonzero affine polynomial on all directions. The Eddington surface boundary condition therefore specifies approximate moments, rather than an exact emergent angular radiation field.

A grey atmosphere has frequency-independent opacity. In plane-parallel radiative equilibrium under the Eddington closure, its temperature obeys

In a well-mixed convective column with fixed absorber mass and constant mass opacity, adding a perfectly transparent gas leaves total optical depth $D$ unchanged. The absolute effective emission pressure scales with total surface pressure, but their ratio remains $D$. In the thick-column emitting-level model, $T_s=T_e{D}^{R_a/c_{pa}}$; added gas warms the surface only if its specific-gas-constant to heat-capacity ratio exceeds that of the original gas. Pressure broadening would violate the fixed-opacity assumption and must be treated separately.

For a source linear in downward effective optical depth $t\in [0,D]$, two-stream approximation weighting $e^{-t}$ has normalized mean $t_{\mathrm{eff}}=1-D/(e^D-1)$. Constant mass opacity gives $p_{\mathrm{rad}}=p_s{t}_{\mathrm{eff}}/D$, tending to $p_s/2$ for small $D$ and $p_s/D$ for large $D$. The atmospheric flux is $(1-e^{-D})S(t_{\mathrm{eff}})$, separate from transmitted ground radiation. Exact angular transfer with literal vertical optical depth instead has weight $2E_2(t)$ and thick-limit mean $2/3$; the convention matters.

A grey atmosphere model with mass absorption coefficient proportional to pressure. Under hydrostatic balance, $\kappa ={\kappa }_{s}p/p_s$ gives downward optical depth $t={\kappa }_s{p}^2/(2g{p}_s)$. For $T=T_s(p/p_s{)}^a$, a non-scattering optically thick column emits hemispheric radiant flux $F\sim \sigma T_s^4{t}_s^{-2a}\Gamma (2a+1)/(a+1)$. The angular factor follows from integrating ray attenuation over the upward hemisphere; a vertical-ray approximation omits it. The extension of the emitting-depth integral to infinity requires fixed $a$ as $t_s\to \infty$.

Consider $I=\alpha [\beta +\tau +\mu +\gamma e^{-\delta \tau }/(1+ϵ\mu )]$ with real $|ϵ|<1$, $\alpha \gamma \ne 0$, and $\delta >0$. Its angular mean includes $A_0=\mathrm{atanh}ϵ/ϵ$. Substitution into $\mu I_{\tau }=I-J$ requires $1-A_0+(\delta -A_0ϵ)\mu =0$. Thus $A_0=1$ and $\delta =ϵ$. For real $ϵ$, $A_0=1$ only at zero, giving $\delta =0$, a contradiction. Independently, its flux contains a varying term proportional to $A_1=(1-A_0)/ϵ$, which is nonzero for nonzero real $ϵ$. Exact conservative transfer therefore needs a more general angular/depth structure than this one smooth exponential correction.

The Hopf function encodes the departure of a plane-parallel grey atmosphere from a purely linear source profile. With physical bolometric radiative flux $F$ and standard mean intensity $J$, it is defined by $J=3F[\tau +q(\tau )]/(4\pi )$. In thermal radiative equilibrium, $T^4=3T_{\mathrm{eff}}^4[\tau +q(\tau )]/4$. The Eddington closure approximation with its approximate surface condition gives constant $q=2/3$; exact angular transfer generally gives a depth-dependent function.

For no incident radiation and outward intensity approximated as direction independent, the surface angular moments are $F=\pi I_0$ and $P_{\mathrm{rad}}(0)=2\pi I_0/(3c)$. Integrating the radiation momentum equation with the Eddington closure approximation yields the grey atmosphere law $T^4=3T_{\mathrm{eff}}^4(\tau +2/3)/4$. This is an approximate angular boundary condition, not isotropy across both hemispheres.

A plane-parallel atmosphere neglects curvature over its vertical thickness, so thermodynamic variables depend on one Cartesian height coordinate.

A plane-parallel atmosphere is geometrically accurate only if the relevant line-forming region is thin relative to the stellar radius. An isothermal estimate gives $H=k_{B}T/(\mu m_u{g}_{\mathrm{eff}})$. Radiative acceleration reduces the effective gravity and increases extension. A small photospheric $H/R$ does not justify a plane-parallel description of wind lines formed far above that photosphere.