Eddington-Barbier Relation

Eddington-Barbier relation is a foundational analytical theorem in astrophysical radiative transfer stating that for a semi-infinite, plane-parallel stellar atmosphere in which the monochromatic source function S_ν(τ_ν) is a linear function of vertical optical depth τ_ν, the emergent specific intensity I_ν⁺(0, μ) observed at direction cosine μ = cos θ equals the source function evaluated at the vertical optical depth τ_ν = μ along the line of sight. Formulated through the radiative equilibrium investigations of British astrophysicist Sir Arthur Stanley Eddington in 1926 and generalized in 1943 by French astronomer Daniel Barbier, the relation provides a direct physical mapping between the angular dependence of a star's surface brightness across its visible disk and the vertical thermal stratification of its photosphere. Across observational astronomy, the Eddington-Barbier relation governs the quantitative interpretation of solar and stellar limb darkening, defines the effective photospheric radius where the emergent astrophysical flux F_ν(0) originates at τ_ν = 2/3, and explains why atomic and molecular transitions appear as dark Fraunhofer absorption lines when temperature decreases outward or as bright emission lines inside chromospheric temperature inversions.

Plane-Parallel Geometry and Optical Depth Coordinates

In the outer photospheric layers of most main-sequence stars, the vertical geometric thickness Δz over which escaping photons are emitted (~300 to 500 km in the Sun) is less than 0.1% of the stellar radius R. Consequently, the stellar atmosphere can be modeled as a horizontally homogeneous, plane-parallel slab stratified along the vertical spatial coordinate z, with the observer located at z → +∞ looking inward at angle θ relative to the outward surface normal. Defining the direction cosine μ = cos θ (where μ = 1 corresponds to the center of the stellar disk and μ → 0 corresponds to the tangential limb) and the vertical monochromatic optical depth dτ_ν = −κ_ν ρ dz measured inward from the top of the atmosphere (τ_ν = 0), the path length element along an inclined ray is ds = dz / cos θ = −dτ_ν / (κ_ν ρ μ).

Substituting this geometric transformation into the time-independent equation of radiative transfer yields the standard plane-parallel transfer equation:

μ (dI_ν / dτ_ν) = I_ν(τ_ν, μ) − S_ν(τ_ν)

where S_ν(τ_ν) = j_ν / κ_ν is the monochromatic source function (the ratio of the mass emission coefficient j_ν to the mass absorption opacity κ_ν, governed at deeper layers by Kramers' opacity law and negative hydrogen H⁻ bound-free continuum absorption). Multiplying both sides by the integrating factor exp(−τ_ν / μ) / μ and integrating from the surface (τ_ν = 0) to infinite depth (τ_ν → ∞, where exp(−τ_ν / μ) I_ν vanishes in a semi-infinite atmosphere) produces the formal integral solution for the emergent outward specific intensity at the stellar surface:

I_ν⁺(0, μ) = ∫₀^∞ S_ν(τ_ν) exp(−τ_ν / μ) (dτ_ν / μ)

Taylor Expansion and the Linear Source Function Theorem

Following Daniel Barbier's 1943 derivation, if the source function S_ν(τ_ν) varies smoothly across the photon escape zone (0.1 ≲ τ_ν ≲ 2), it can be expanded as a polynomial or Taylor series around τ_ν = 0:

S_ν(τ_ν) = a_ν + b_ν τ_ν + (c_ν / 2) τ_ν² + …

Recalling the standard Euler gamma-function integral identity ∫₀^∞ xⁿ exp(−x) dx = n! with the substitution x = τ_ν / μ (so τ_ν = μ x and dτ_ν / μ = dx), each power-law term integrates analytically:

I_ν⁺(0, μ) = a_ν ∫₀^∞ exp(−x) dx + b_ν μ ∫₀^∞ x exp(−x) dx + (c_ν / 2) μ² ∫₀^∞ x² exp(−x) dx + …

I_ν⁺(0, μ) = a_ν + b_ν μ + c_ν μ² + …

In a gray atmosphere under the Eddington approximation or any local thermodynamic equilibrium (LTE) photosphere where higher-order quadratic curvature is small (c_ν ≈ 0), the source function is linear in optical depth: S_ν(τ_ν) = a_ν + b_ν τ_ν. Comparing the functional form of S_ν(τ_ν) with the integrated surface intensity I_ν⁺(0, μ) = a_ν + b_ν μ yields the Eddington-Barbier relation for specific intensity:

I_ν⁺(0, μ) = S_ν(τ_ν = μ)

Physically, a photon escaping the atmosphere at viewing angle θ travels one slant optical depth unity (Δτ_slant = τ_ν / μ = 1) before escaping to infinity; projected onto the vertical axis, this unit slant optical depth corresponds to the vertical optical depth τ_ν = μ = cos θ.

Emergent Flux and the τ = 2/3 Photospheric Surface

For unresolved stars beyond the Sun, telescopes measure the disk-integrated emergent monochromatic flux F_ν⁺(0) rather than spatially resolved intensity I_ν⁺(0, μ):

F_ν⁺(0) = 2π ∫₀¹ I_ν⁺(0, μ) μ dμ

Substituting the linear Eddington-Barbier intensity I_ν⁺(0, μ) = a_ν + b_ν μ into the flux integral gives:

F_ν⁺(0) = 2π ∫₀¹ (a_ν μ + b_ν μ²) dμ = 2π [ (a_ν / 2) + (b_ν / 3) ] = π [ a_ν + (2/3) b_ν ]

Recognizing that a_ν + (2/3) b_ν is the linear source function evaluated at τ_ν = 2/3 yields the Eddington-Barbier flux relation:

F_ν⁺(0) = π S_ν(τ_ν = 2/3)

Because a blackbody of effective temperature T_eff radiates a bolometric surface flux F = σ_SB T_eff⁴ = π B(T_eff), the Eddington-Barbier flux theorem proves that a star's effective temperature T_eff—the horizontal coordinate of the Hertzsprung–Russell diagram and the baseline of Strömgren photometry and the Baade-Wesselink method—equals the local kinetic gas temperature at vertical Rosseland optical depth τ_R = 2/3.

Solar Limb Darkening and Empirical Temperature Inversions

When observing the solar disk in visible light, the center of the disk (θ = 0°, μ = 1) appears substantially brighter and bluer than the edge or limb of the disk (θ → 90°, μ → 0). Under the Eddington-Barbier relation I_ν⁺(0, μ) = S_ν(τ_ν = μ) ≈ B_ν[T(τ_ν = μ)], lines of sight near disk center penetrate vertically down to deeper, hotter layers at τ_ν = 1 (T ≈ 6,400 K in the solar photosphere), whereas grazing lines of sight near the solar limb reach slant optical depth unity within shallow, cooler upper photospheric layers at τ_ν = μ ≪ 1 (T ≈ 4,500 K near the temperature minimum).

Normalized to disk center (I_ν⁺(0, 1) = a_ν + b_ν), the linear limb-darkening law takes the form:

I_ν⁺(0, μ) / I_ν⁺(0, 1) = 1 − u_ν (1 − μ)

where u_ν = b_ν / (a_ν + b_ν) is the monochromatic linear limb-darkening coefficient. Accurate limb-darkening parameterizations derived from the Eddington-Barbier expansion are essential for modeling exoplanet transit light curves, eclipsing binary photometry during solar eclipses, and rotational obliquity waveforms via the Rossiter–McLaughlin effect.

Formation of Absorption vs. Emission Spectral Lines

The Eddington-Barbier relation provides an immediate physical explanation for why stellar spectra display dark absorption lines or bright emission lines across a fixed line of sight (for example, at disk center μ = 1). Across an atomic spectral transition centered at frequency ν₀, the total opacity is the sum of the smooth background continuum opacity κ_cont and the sharply peaked line opacity κ_line(ν). Because the Eddington-Barbier relation dictates that escaping photons always originate where the total monochromatic optical depth equals unity (τ_total,ν = 1):

1. In the Adjacent Continuum (ν ≠ ν₀): Opacity is low (κ_ν = κ_cont), so τ_cont = 1 is reached deep in the photosphere where the gas temperature T(z_cont) is high.
2. In the Spectral Line Core (ν = ν₀): Opacity is orders of magnitude higher (κ_ν = κ_cont + κ_line ≫ κ_cont), so total optical depth unity (τ_line = 1) is reached at a much higher geometric altitude z_line > z_cont where the continuum optical depth is tiny: τ_cont(z_line) = κ_cont / (κ_cont + κ_line) ≪ 1.

In a normal stellar photosphere heated from below by nuclear fusion (at the Gamow peak) and Kelvin-Helmholtz contraction, temperature decreases monotonically with altitude (T(z_line) < T(z_cont), or dT/dτ > 0). By the Eddington-Barbier relation under LTE (S_ν = B_ν(T)), the line-core intensity I_ν0(0, 1) = B_ν0[T(z_line)] is lower than the continuum intensity I_cont(0, 1) = B_ν0[T(z_cont)], producing a dark Fraunhofer absorption line. Conversely, if acoustic and magnetohydrodynamic wave dissipation heats an overlying chromosphere or Parker solar wind base so that temperature increases outward (T(z_line) > T(z_cont)), the Eddington-Barbier relation predicts I_ν0 > I_cont, generating an emission line (or self-reversed emission peaks in the cores of Ca II H & K and Mg II h & k lines) as well as limb brightening at ultraviolet, millimeter, and radio wavelengths.

Although the Eddington-Barbier relation is exact for linear source functions in plane-parallel geometry, two important departures arise in modern radiative transfer modeling:

• Spherical Extension in Supergiants and Stellar Winds: In red supergiants, asymptotic giant branch stars, and dense stellar winds of Wolf-Rayet stars and Luminous Blue Variables, the atmospheric scale height H is comparable to the stellar radius R. Spherical divergence invalidates the plane-parallel coordinate transformation ds = −dτ / (κ ρ μ), causing extended scattering halos and wavelength-dependent photospheric radii.
• Non-LTE Resonance Scattering (Two-Level Atom): High in the chromosphere where gas density ρ is low, radiative de-excitation rates A₂₁ far exceed collisional thermalization rates C₂₁ (ε = C₂₁ / [C₂₁ + A₂₁] ≪ 1). The line source function decouples from the local Planck function B_ν(T) and relaxes toward the mean scattering intensity J_ν, scaling near the surface as S_ν(0) ≈ √ε B_ν(T) ≪ B_ν(T). Even when a chromospheric temperature inversion is present, resonance scattering drives deep dark cores in strong metal lines while still satisfying the generalized integral weighting of the Eddington-Barbier contribution function.

1. Eddington, A. S. (1926). The Internal Constitution of the Stars. Cambridge University Press. https://doi.org/10.1017/CBO9780511600005
2. Barbier, D. (1943). "Sur la théorie du spectre continu des étoiles." Annales d'Astrophysique, 6, 113–133. https://ui.adsabs.harvard.edu/abs/1943AnAp....6..113B
3. Rutten, R. J. (2003). Radiative Transfer in Stellar Atmospheres (8th ed.). Utrecht University Lecture Notes. https://ui.adsabs.harvard.edu/abs/2003rtsa.book.....R
4. Hubeny, I., & Mihalas, D. (2014). Theory of Stellar Atmospheres: An Introduction to Astrophysical Non-equilibrium Quantitative Spectroscopic Analysis. Princeton University Press. https://doi.org/10.1515/9781400852734