Interstellar Extinction and Reddening

Why interstellar extinction matters

Almost nothing in the Galaxy is seen in its true colors. Between us and every distant star lies a thin haze of solid particles — the dust component of the interstellar medium — that intercepts a fraction of the light. Ignore it, and a star looks fainter than its intrinsic luminosity and redder than its true temperature. Both errors propagate straight into the most fundamental measurements in astronomy.

• Distances. Standard candles such as Cepheids and Type Ia supernovae work only if the observed brightness reflects the true brightness. Uncorrected extinction makes objects seem farther away; a 1-magnitude error in A_V translates into a ~60% error in inferred distance.
• Stellar temperatures. Colors are used to estimate effective temperature. Reddening mimics a cooler star, biasing spectral typing and the placement of stars on the Hertzsprung–Russell diagram.
• The dust itself. The extinction curve is our primary probe of grain size, composition, and abundance — dust is the raw material of planets and the surface where H₂ and complex molecules form.
• Cosmology. Both foreground Milky Way dust and dust in host galaxies redden supernovae; disentangling dust reddening from cosmological dimming is central to measuring the expansion history and the Hubble constant.
• Mapping the Galaxy. Extinction traces the spiral arms and dense clouds; combined with parallaxes from Gaia, it yields three-dimensional dust maps of the Milky Way.

How extinction works, step by step

Extinction is the combined effect of two processes acting on photons traveling through a dusty medium:

1. Absorption. A grain absorbs a photon, converting its energy into heat. The grain re-radiates that energy in the far-infrared, so absorbed starlight is not destroyed but shifted to wavelengths far from where it started.
2. Scattering. A grain deflects a photon out of the line of sight without absorbing it. That photon is not lost to the Universe — it may be seen as a reflection nebula — but it is lost to our particular beam, so it still counts as extinction.

The efficiency of both processes depends on the ratio of grain size to wavelength. Interstellar grains span roughly 0.01–0.3 µm, straddling the wavelengths of visible light. In this regime the scattering cross-section rises steeply toward the blue, roughly as 1/λ across the optical (softer than the 1/λ⁴ of true Rayleigh scattering, because the grains are not point-like). More blue photons are removed than red, so the transmitted light is both dimmer and redder. The same selective scattering reddens the setting Sun and paints the daytime sky blue.

The amount of light surviving falls off exponentially with the optical depth τ(λ): a star's flux is reduced by e^−τ(λ). Astronomers convert this into magnitudes, giving the extinction

A(λ) = m(λ)_observed − m(λ)_intrinsic = 1.086 × τ(λ)

where the factor 1.086 = 2.5/ln(10) links optical depth to the magnitude system. The wavelength-dependent function A(λ), normalized in some way, is the extinction curve.

Reddening, color excess, and R_V

Because A(λ) is larger in the blue than the red, the extinction in the B band exceeds that in the V band. The difference is the color excess:

E(B−V) = A_B − A_V = (B−V)_observed − (B−V)_intrinsic

E(B−V) is directly measurable if you know the star's intrinsic color (from its spectral type), so it is the workhorse quantity for quantifying reddening. To convert reddening into total extinction we use the ratio of total-to-selective extinction:

R_V = A_V / E(B−V)

R_V encodes the slope of the extinction across the optical — physically, the mean grain size. Small grains produce steep reddening and a low R_V; large grains give flatter, "grayer" extinction and a high R_V. The diffuse Milky Way averages R_V ≈ 3.1, so a common rule of thumb is A_V ≈ 3.1 × E(B−V). Dense molecular clouds, where grains have grown by coagulation and ice mantles, can reach R_V = 4–6.

Extinction in common photometric bands relative to A_V (Milky Way, R_V = 3.1)

Band	Central λ	A(band) / A_V	Comment
U (ultraviolet)	0.365 µm	≈ 1.53	Strongly extinguished
B (blue)	0.44 µm	≈ 1.32	A_B = A_V + E(B−V)
V (visual)	0.55 µm	1.00	Reference band
R (red)	0.70 µm	≈ 0.75	Less dimmed
I (near-IR)	0.80 µm	≈ 0.59	About 60% of the visual extinction
J	1.25 µm	≈ 0.28	Infrared window opens
H	1.65 µm	≈ 0.18	Dust becoming transparent
K	2.2 µm	≈ 0.11	Sees through the Galactic center

The extinction curve and the 2175 Å bump

Plot A(λ)/A(V) — or the more common A(λ)/E(B−V) — against 1/λ and you get the interstellar extinction curve. Moving from the infrared to the ultraviolet it does the following: it rises gently through the optical; it carries a broad, symmetric absorption feature centered at 2175 Å (0.2175 µm) in the ultraviolet; then it climbs steeply into the far-UV below about 1500 Å. The near-IR portion follows a smooth power law, A(λ) ∝ λ^−1.6 or so.

The 2175 Å bump was discovered by Theodore Stecher in 1965 using rocket-borne UV photometry. Its central wavelength is astonishingly stable from sight line to sight line, but its width varies. The consensus carrier is small carbonaceous grains — polycyclic aromatic hydrocarbons (PAHs) and graphitic or amorphous-carbon particles — absorbing via a π→π* electronic transition. Crucially, the bump is prominent in the Milky Way, weaker in the Large Magellanic Cloud, and essentially absent along most sight lines in the Small Magellanic Cloud, telling us that dust chemistry and processing differ between galaxies.

Because the whole curve — infrared power law, optical slope, UV rise, and bump strength — correlates with a single parameter, the Cardelli, Clayton & Mathis (1989) parameterization famously expresses A(λ)/A_V as a function of wavelength that depends only on R_V. The Fitzpatrick (1999) law and O'Donnell (1994) refinements are widely used variants. This is enormously convenient: measure one number, R_V, and you predict extinction at every wavelength.

Worked example: dereddening a reddened B star

Suppose spectroscopy identifies a main-sequence B0 V star, whose intrinsic color is (B−V)₀ = −0.30. You observe it at (B−V) = +0.15. Then:

• Color excess: E(B−V) = (B−V) − (B−V)₀ = 0.15 − (−0.30) = 0.45 mag.
• Total visual extinction (assuming diffuse dust, R_V = 3.1): A_V = R_V × E(B−V) = 3.1 × 0.45 ≈ 1.40 mag. The star is dimmed in V by a factor of 10^0.56 ≈ 3.6.
• Dereddened magnitude: V₀ = V_obs − A_V. If V_obs = 11.0, then V₀ ≈ 9.6.
• Distance: now apply the distance modulus m − M = 5 log₁₀(d/10 pc), using V₀ and the star's absolute magnitude. Had you skipped the extinction correction, you would have used V_obs and overestimated the distance by nearly a factor of two.

This is exactly why dust maps matter. The Schlegel, Finkbeiner & Davis (1998) all-sky map, and more recently Planck and three-dimensional Gaia-based maps, let you look up E(B−V) toward any position and correct photometry even without a spectral type.

Common misconceptions

• "Extinction destroys the light." No — absorbed energy is re-emitted in the far-infrared, and scattered light reappears elsewhere (reflection nebulae). Extinction removes light from your specific line of sight, not from the Universe.
• "Reddening means the star is redshifted." Unrelated. Redshift stretches wavelengths uniformly; reddening changes the shape of the spectrum by preferentially removing blue light. A reddened star's spectral lines sit at their normal wavelengths.
• "All sight lines follow R_V = 3.1." That is only the diffuse-ISM average. Dense clouds reach R_V = 4–6, and using 3.1 there under-corrects the extinction.
• "Infrared light passes through dust untouched." It is far less affected — A_K ≈ 0.11 A_V — but not immune, and silicate features return near 10 µm.
• "Extinction and dust column are the same thing." A_V traces the dust column, which correlates with the gas column (roughly N_H ≈ 5.8×10²¹ × E(B−V) cm⁻², or N_H/A_V ≈ 1.9×10²¹ cm⁻² mag⁻¹), but the conversion depends on the gas-to-dust ratio, which varies with metallicity.
• "Gray extinction cannot happen." If grains grow large enough, extinction becomes nearly wavelength-independent — dimming without reddening — which is precisely the worry for supernova cosmology.