Interstellar Extinction and Reddening

Look at the Milky Way on a dark night and you see it split lengthwise by a ragged dark lane. That lane is not a gap between stars — it is dust, absorbing and scattering the light of everything behind it. Every photon that reaches a telescope from a distant star has run this gauntlet of submicron grains, and it arrives both fainter and redder than it left. The dimming is extinction; the colour change is reddening; and because the grains are comparable in size to a wavelength of visible light, the two are inseparable manifestations of one process that depends steeply on wavelength — blue light is removed far more efficiently than red. This article assembles the machinery an observer actually uses. Extinction is bookkept in magnitudes, $A_\lambda=-2.5\log_{10}(F_\lambda/F_{\lambda,0})$ , which makes it additive and ties it to optical depth through the universal factor $A_\lambda=1.086\,\tau_\lambda$ . The wavelength dependence is compressed into two numbers: the color excess $E(B-V)$ that measures how much redder a star looks, and the ratio of total to selective extinction $R_V\equiv A_V/E(B-V)$ , whose diffuse-ISM value near $3.1$ is one of the most quoted constants in Galactic astronomy. Remarkably, Cardelli, Clayton and Mathis showed that the whole extinction curve from the infrared to the ultraviolet — power-law IR rise, the 2175 Å bump, and the steep far-UV climb — collapses onto a one-parameter family in $R_V$ alone. Correcting for all of this is a prerequisite for every distance, luminosity, stellar age, and — at high Galactic latitude — every cosmological measurement built on standard candles.

Key takeaways

Interstellar dust grains scatter and absorb starlight, dimming a source by $A_\lambda$ magnitudes that grows toward the blue, so stars look both fainter and redder than their true colours.
Extinction is additive in magnitudes because it is logarithmic in flux: $A_\lambda=1.086\,\tau_\lambda$ , with $1.086=2.5/\ln 10$ linking magnitudes to optical depth.
Reddening is the color excess $E(B-V)=A_B-A_V$ ; total dimming is $A_V$ ; their ratio $R_V\equiv A_V/E(B-V)\approx3.1$ in the diffuse ISM and rises to $4$ – $6$ in dense clouds where grains have grown.
The full curve $A_\lambda/A_V$ versus $1/\lambda$ is a one-parameter family in $R_V$ (Cardelli, Clayton and Mathis 1989); it carries the 2175 Å bump from small carbonaceous grains and a steep far-UV rise from the smallest grains.
Dust traces gas: $N_\mathrm{H}/E(B-V)\approx5.8\times10^{21}\ \mathrm{cm^{-2}\,mag^{-1}}$ , so reddening maps double as hydrogen-column maps.
Key idea: dust removes light wavelength-selectively — total dimming $A_V$ and reddening $E(B-V)$ are locked together by $R_V=A_V/E(B-V)\approx3.1$ , and the extinction curve $A_\lambda/A_V$ is a compact fingerprint of grain size and composition.

The core insight

Dust removes light wavelength-selectively: total dimming $A_V$ and the reddening $E(B-V)$ are tied together by $R_V=A_V/E(B-V)\approx3.1$ , and the full extinction curve $A_\lambda/A_V$ is a compact fingerprint of grain size and composition.

Where this lives today

Extinction correction has moved from a nuisance factor to a precision science with its own instruments. Three-dimensional dust maps — built by combining stellar photometry and parallaxes from Gaia with near-infrared surveys — now give $E(B-V)$ as a function of distance along millions of sightlines, so an astronomer can deredden a target by looking up the dust in front of it rather than assuming a smooth screen. All-sky two-dimensional maps from far-infrared dust emission (Schlegel, Finkbeiner and Davis 1998, and its Planck-era successors) supply the Galactic foreground reddening subtracted from every extragalactic measurement. That subtraction is not academic for cosmology: Type Ia supernova distances, the backbone of cosmic acceleration, depend on dereddening both the Galactic foreground and the host-galaxy dust, and the assumed $R_V$ toward supernovae remains an active systematic in dark-energy analyses. At the frontier of grain physics, JWST is dissecting extinction and ice features toward embedded protostars, while the identity of the 2175 Å bump carrier and the origin of the diffuse interstellar bands are still open problems. The magnitude of dimming that Trumpler first measured now propagates into the equation of state of dark energy.

The physical picture

Picture starlight entering a thin fog whose droplets happen to be about the size of a wavelength of visible light. Each grain intercepts photons — some absorbed and converted to heat, some scattered out of the beam — and the efficiency of that interception climbs steeply toward short wavelengths. Blue light is therefore stripped from the beam far more aggressively than red, exactly as a distant streetlight in smog fades to a dim orange. The star is not moved along the spectrum the way a redshift would move it; its whole continuum is tilted toward the red and lowered in brightness. The first figure fixes this two-part effect — a dimming and a colour tilt — with self-consistent numbers.

Figure

Starlight crossing a dust column is dimmed and reddened: blue suffers A_B = 2.05 mag of extinction, visual A_V = 1.55 mag, so the color excess is E(B−V) = A_B − A_V = 0.50 mag and the total-to-selective ratio is R_V = A_V ÷ E(B−V) = 3.1

The reason blue is hit harder is that a grain of radius $a$ presents an effective cross-section $\sigma_\lambda=\pi a^2 Q_\mathrm{ext}(\lambda)$ whose efficiency $Q_\mathrm{ext}$ depends on the size parameter $x=2\pi a/\lambda$ . When the grain is much smaller than the wavelength it scatters weakly and pure absorption dominates; when it is comparable to the wavelength the interception is strongest and richly structured; when it is much larger the efficiency saturates and the extinction goes "grey", dimming all colours alike. A real sightline mixes a broad range of grain sizes, and the sum of their efficiencies traces out the observed extinction curve. That curve — the single most information-rich object in the whole subject — is the punchline of the second figure, plotted the way astronomers always plot it, against inverse wavelength so that the ultraviolet is on the right.

Figure

The Galactic extinction curve A_λ ÷ A_V versus inverse wavelength x = 1÷λ for the diffuse-ISM value R_V = 3.1: it rises gently from the near-IR, passes through V (x = 1.82, ratio 1) and B (x = 2.27, ratio 1.32), peaks at the 2175 Å carbonaceous bump (x = 4.6), and climbs steeply into the far-UV

Two structural facts organize everything below. First, $R_V$ is a proxy for grain size: larger grains scatter more greyly, flattening the optical slope and pushing $R_V$ up, so a large $R_V$ toward a molecular cloud is telling you the grains have coagulated and grown ice mantles. Second, reddening and extinction always arrive as a pair because one grain population causes both — measure a star's intrinsic colour from its spectral type, read off the color excess, and multiply by $R_V$ to recover the total dimming. The rest of the article turns these two facts into equations, derivations, and a working dereddening recipe.

The core equations

Extinction is defined as a difference of magnitudes, which is the same as $-2.5\log_{10}$ of the ratio of observed to intrinsic flux. Because it is logarithmic, it is non-negative for a source seen through dust and it simply adds when light crosses several clouds in succession.

A_\lambda=-2.5\log_{10}\!\left(\frac{F_\lambda}{F_{\lambda,0}}\right)=m_\lambda-m_{\lambda,0} \qquad (1)

where: $A_\lambda$ — extinction at wavelength $\lambda$ , in magnitudes ( $\ge 0$ ); $F_\lambda,\ F_{\lambda,0}$ — observed and intrinsic (unextinguished) fluxes; $m_\lambda,\ m_{\lambda,0}$ — observed and intrinsic apparent magnitudes. Reading: $A_\lambda$ is just how many magnitudes of the source's light the dust ate at wavelength $\lambda$ . The physics under that definition is exponential attenuation. Light crossing dust of optical depth $\tau_\lambda$ is transmitted as $e^{-\tau_\lambda}$ , and feeding that into equation (1) produces a fixed conversion constant between optical depth and magnitudes.

A_\lambda=1.086\,\tau_\lambda,\qquad \tau_\lambda=\int_0^s n_d(s')\,\sigma_\lambda\,\mathrm{d}s' \qquad (2)

where: $\tau_\lambda$ — dust optical depth along the sightline; $n_d$ — number density of grains; $\sigma_\lambda=\pi a^2 Q_\mathrm{ext}(\lambda)$ — extinction cross-section per grain; $1.086=2.5/\ln 10$ . Reading: one magnitude of extinction is very nearly one optical depth ( $\tau=1$ gives $A=1.086$ ). Because $A_\lambda$ falls with wavelength, the blue band is dimmed more than the visual, and the difference is a positive color excess — the observed color index minus the intrinsic one. This is what "reddening" means operationally: a change in colour, not a change in brightness.

E(B-V)\equiv (B-V)_\mathrm{obs}-(B-V)_0 = A_B-A_V \qquad (3)

where: $(B-V)_\mathrm{obs}$ — observed color index; $(B-V)_0$ — intrinsic color, from spectral type or a calibrated relation; $A_B,\ A_V$ — extinction in the $B$ and $V$ bands. Reading: $E(B-V)$ is the reddening you can measure without ever knowing the star's distance. The bridge from the colour change to the total dimming is the single most important derived constant in the field: the ratio of total ( $A_V$ ) to selective ( $E(B-V)$ ) extinction. It encodes the optical slope of the extinction curve and correlates with mean grain size.

R_V\equiv\frac{A_V}{E(B-V)}\approx 3.1\quad\text{(diffuse ISM)} \qquad (4)

where: $R_V$ — total-to-selective extinction ratio, larger for larger grains; $A_V$ — total visual extinction; $E(B-V)$ — color excess. Reading: multiply a measured reddening by $R_V$ and you have the actual number of magnitudes lost in $V$ . The astonishing empirical fact is that this one number fixes the entire curve. Cardelli, Clayton and Mathis parametrized $A_\lambda/A_V$ from the near-infrared to the ultraviolet as a fixed shape plus an $R_V$ -dependent correction.

\frac{A_\lambda}{A_V}=a(x)+\frac{b(x)}{R_V},\qquad x=\frac{1}{\lambda\,[\mu\mathrm{m}]} \qquad (5)

where: $a(x),\ b(x)$ — universal polynomial functions of inverse wavelength; $x$ — inverse wavelength in $\mu\mathrm{m}^{-1}$ ; $R_V$ — the sole free parameter. Reading: fix $R_V$ and you have predicted the extinction at every wavelength, which is why dereddening in practice needs only $E(B-V)$ and an assumed $R_V$ . Finally, dust and gas are well mixed, so the reddening also measures a hydrogen column — the relation that turns optical reddening maps into gas-column maps.

\frac{N_\mathrm{H}}{E(B-V)}\approx 5.8\times10^{21}\ \mathrm{atoms\,cm^{-2}\,mag^{-1}} \qquad (6)

where: $N_\mathrm{H}$ — total hydrogen column, H I plus twice H $_2$ ; $E(B-V)$ — color excess. Reading: divide by $R_V\approx3.1$ to get $N_\mathrm{H}/A_V\approx1.9\times10^{21}\ \mathrm{cm^{-2}\,mag^{-1}}$ , the form used to weigh molecular clouds from their extinction.

Derivation AWhy extinction equals 1.086 times the optical depth

The goal is to explain the ubiquitous factor $1.086$ that converts optical depth into magnitudes, and to show why this same logarithmic structure is what makes extinction additive across independent dust clouds — a property that is not obvious from the flux, which multiplies, but is essential to every dereddening calculation in astronomy. We start from radiative transfer through a purely attenuating medium, insert it into the magnitude definition, and read off the constant.

Step 1. For a beam that is only absorbed and scattered out of the line of sight, radiative transfer gives $F_\lambda=F_{\lambda,0}\,e^{-\tau_\lambda}$ , where $\tau_\lambda$ is the dust optical depth. Why: with no emission or in-scattering back into the beam, each layer removes a fixed fraction of the remaining photons, and integrating that fractional loss exponentiates the optical depth.

Step 2. Substitute into the definition $A_\lambda=-2.5\log_{10}(F_\lambda/F_{\lambda,0})$ to get $A_\lambda=-2.5\log_{10}(e^{-\tau_\lambda})=2.5\,\tau_\lambda\log_{10}e$ . Why: the magnitude system is base-10 logarithmic while the physics is base- $e$ , so converting between them is the entire content of the numerical factor about to appear.

Step 3. Evaluate $\log_{10}e=1/\ln 10=0.4343$ , giving $A_\lambda=(2.5/\ln 10)\,\tau_\lambda=1.086\,\tau_\lambda$ . Why: the constant is exact — $2.5\times0.4343=1.0857$ — so $\tau=1$ corresponds to $A=1.086$ mag, and $A_\lambda$ is essentially the optical depth measured in magnitude units.

Step 4. Additivity follows immediately: two clouds with optical depths $\tau_1$ and $\tau_2$ transmit $e^{-\tau_1}e^{-\tau_2}=e^{-(\tau_1+\tau_2)}$ , so $A_\mathrm{tot}=1.086(\tau_1+\tau_2)=A_1+A_2$ . Why: this is why extinction is quoted in magnitudes at all — dust columns bookkeep like a running sum, and a foreground and a background cloud simply add their $A_\lambda$ contributions.

Go deeper · Derivation B · 2 minFrom grain cross-sections to the extinction curve and R_V

Derivation BFrom grain cross-sections to the extinction curve and R_V

The goal is to build the observed extinction curve from microscopic grain physics — a size distribution of particles each with a wavelength-dependent efficiency — and to show precisely why the optical slope of that curve defines $R_V$ and makes it a proxy for grain size. We integrate the single-grain optical depth over both the path and the grain-size distribution, track how the efficiency $Q_\mathrm{ext}$ behaves in the three size regimes, and identify the combination that observers call $R_V$ .

Step 1. Each grain removes photons with cross-section $\sigma_\lambda=\pi a^2 Q_\mathrm{ext}(a,\lambda)$ , so a thin layer contributes $\mathrm{d}\tau_\lambda=n_d\,\sigma_\lambda\,\mathrm{d}s$ . Why: the geometric area $\pi a^2$ sets the scale, and the dimensionless efficiency $Q_\mathrm{ext}$ carries all the wavelength dependence through the size parameter $x=2\pi a/\lambda$ .

Step 2. Integrate over the path and the grain-size distribution $n(a,s')$ :

A_\lambda=1.086\int_0^s\!\!\int_{a_-}^{a_+}\pi a^2\,Q_\mathrm{ext}(a,\lambda)\,n(a,s')\,\mathrm{d}a\,\mathrm{d}s'

The classic MRN model takes a power law $n(a)\propto a^{-3.5}$ over $0.005\lesssim a\lesssim0.25\ \mu\mathrm{m}$ . Why: no single grain size reproduces the curve; the steep power law puts most of the geometric surface area in the smallest grains, which is what supplies the ultraviolet extinction.

Step 3. The wavelength dependence enters entirely through $Q_\mathrm{ext}(x)$ : for $x\ll1$ (grains much smaller than $\lambda$ ) $Q_\mathrm{ext}\propto\lambda^{-4}$ and pure absorption dominates; for $x\sim1$ the efficiency is largest and structured; for $x\gg1$ it saturates at $Q_\mathrm{ext}\to2$ , giving grey extinction. Why: the far-UV rise comes from small grains reaching their $x\sim1$ peak at short wavelengths, while big grains dim all colours equally — the mixture sets the overall slope.

Step 4. Define the optical slope through the visual and blue bands as $R_V=A_V/(A_B-A_V)$ . A size distribution weighted toward large grains flattens the $A_B$ -to- $A_V$ contrast, lowering $A_B-A_V$ at fixed $A_V$ and so raising $R_V$ . Why: this is the precise sense in which $R_V$ measures mean grain size — greyer (flatter) extinction means bigger grains means larger $R_V$ .

Step 5. Cardelli, Clayton and Mathis found empirically that the whole curve collapses to $A_\lambda/A_V=a(x)+b(x)/R_V$ , a one-parameter family. Why: fixing the optical slope $R_V$ turns out to fix the infrared power law and the ultraviolet rise too, so a single measured number specifies the entire dereddening law — the regularity that makes practical extinction correction possible.

Go deeper · Derivation C · 1 minTurning a reddened star's colours into its true distance

Derivation CTurning a reddened star's colours into its true distance

The goal is the observer's core workflow: given a star of known spectral type, use its intrinsic colour and absolute magnitude together with its observed photometry to extract the reddening, the total extinction, and finally an extinction-corrected distance — and to make explicit how badly the distance is biased if extinction is ignored. We proceed from the measured colour to the color excess, apply $R_V$ , and correct the distance modulus.

Step 1. The spectral type fixes the intrinsic colour $(B-V)_0$ and absolute magnitude $M_V$ from a calibration; measure $(B-V)_\mathrm{obs}=B_\mathrm{obs}-V_\mathrm{obs}$ directly. Why: the star's own spectrum, through its type, supplies the "unreddened template" that the dust is compared against — no distance is needed for this step.

Step 2. Form the color excess $E(B-V)=(B-V)_\mathrm{obs}-(B-V)_0$ . Why: this isolates the purely differential effect of dust, cancelling the star's intrinsic colour so that what remains is the reddening alone.

Step 3. Convert to total extinction with $A_V=R_V\,E(B-V)\approx3.1\,E(B-V)$ . Why: $R_V$ is the only bridge from the colour change you measured to the brightness loss you need; for diffuse sightlines $3.1$ is safe, but toward clouds $R_V$ must be checked, not assumed.

Step 4. Correct the distance modulus, $m_V-M_V=5\log_{10}(d/10\,\mathrm{pc})+A_V$ , and solve $d=10^{(m_V-M_V-A_V+5)/5}\ \mathrm{pc}$ . Why: extinction masquerades as extra distance, so it must be subtracted from the apparent modulus before inverting — omitting a real $A_V$ inflates the inferred distance by the multiplicative factor $10^{A_V/5}$ , which is already $1.58\times$ for a single magnitude. *The distance bias factor $10^{A_V/5}$ is exact.*

Limiting cases

The physics is governed by the grain size relative to wavelength, captured by the size parameter $x=2\pi a/\lambda$ . The three regimes below explain, respectively, the flat infrared tail, the structured optical curve, and the grey extinction of large grains.

Case	Result	Meaning
Rayleigh limit ( $x\ll1$ , small grains or long $\lambda$ )	$Q_\mathrm{ext}\propto\lambda^{-4}$	Cross-section collapses toward long wavelengths, so far-IR extinction is small and nearly grey; scattering is negligible and absorption dominates.
Resonance regime ( $x\sim1$ , optical)	$A_\lambda\propto\lambda^{-1}$ (approx.)	Grain and wavelength are matched; the extinction is strongest and most structured, and the roughly $\lambda^{-1}$ optical slope and 2175 Å bump live here.
Geometric-optics limit ( $x\gg1$ , large grains or short $\lambda$ )	$Q_\mathrm{ext}\to 2$	Efficiency saturates at twice the geometric cross-section (the extinction paradox), giving grey extinction that dims all colours equally and raises $R_V$ .

Common trapReddening is not a redshift, and reddening is not the same as extinction

Three confusions cost real errors. First, reddening (a colour change, $E(B-V)$ ) and extinction (a brightness change, $A_V$ ) are different quantities — you cannot get one from the other without $R_V$ , and treating $E(B-V)$ as if it were $A_V$ underestimates the dimming by a factor of about three. Second, reddening is not a redshift: dust tilts the continuum by removing more blue than red light, but every spectral line stays put at its rest or Doppler wavelength, whereas a redshift slides the entire spectrum bodily along the wavelength axis. Third, dust does not add emission lines — it removes and scatters continuum photons; any emission you see is from the gas or from the dust's own thermal re-radiation in the far-infrared, a separate process.

Worked examples

1 · From colour to dimming to band extinctions (confidence builder)

A star is observed at $(B-V)_\mathrm{obs}=0.8$ , but its spectral type implies an intrinsic $(B-V)_0=0.2$ . Assuming the diffuse-ISM value $R_V=3.1$ , find the reddening, the visual extinction, and the extinction in the $B$ band, and check the implied $R_B$ .

Step 1. The color excess is just the difference of colours: $E(B-V)=(B-V)_\mathrm{obs}-(B-V)_0=0.8-0.2=\mathbf{0.6\ mag}$ . The star looks $0.6$ mag redder in $B-V$ than it truly is.

Step 2. Convert to total visual extinction with $R_V$ : $A_V=R_V\,E(B-V)=3.1\times0.6=\mathbf{1.86\ mag}$ . So the star is dimmed by nearly two magnitudes — a factor $10^{1.86/2.5}=5.5$ in $V$ -band flux — purely by dust.

Step 3. The $B$ -band extinction follows from the definition of the color excess, $E(B-V)=A_B-A_V$ , so $A_B=A_V+E(B-V)=1.86+0.6=\mathbf{2.46\ mag}$ . Blue light is dimmed more than visual, as it must be.

Step 4. As an internal check, the blue total-to-selective ratio is $R_B\equiv A_B/E(B-V)=2.46/0.6=\mathbf{4.1}$ , which is exactly $R_V+1$ . E(B−V) = 0.6, A_V = 1.86, A_B = 2.46, and R_B = R_V + 1 = 4.1 — the tidy relation $R_B=R_V+1$ is a good sanity test of any band-extinction bookkeeping.

2 · The distance bias from ignoring extinction (fully worked)

A reddened B0 V star has calibrated intrinsic values $(B-V)_0=-0.30$ and $M_V=-4.1$ . It is observed at $V=7.0$ and $B=7.1$ . Take $R_V=3.1$ . Find its true distance, and quantify how far wrong you would be if you neglected extinction. Also estimate the hydrogen column toward it.

Step 1. Observed colour: $(B-V)_\mathrm{obs}=B-V=7.1-7.0=0.10$ . Color excess: $E(B-V)=0.10-(-0.30)=\mathbf{0.40\ mag}$ . The intrinsically blue B star has been reddened to nearly neutral.

Step 2. Total extinction: $A_V=R_V\,E(B-V)=3.1\times0.40=\mathbf{1.24\ mag}$ . This is the number that must come out of the distance modulus.

Step 3. Apparent modulus: $m_V-M_V=7.0-(-4.1)=11.1$ . Subtract extinction for the true modulus $11.1-1.24=9.86$ , giving $d=10^{(9.86+5)/5}=10^{2.972}=\mathbf{938\ pc}$ .

Step 4. Neglecting extinction entirely would give $d_\mathrm{wrong}=10^{(11.1+5)/5}=10^{3.22}=1660$ pc. The ratio is $10^{A_V/5}=10^{0.248}=1.77$ , so ignoring $A_V=1.24$ mag overestimates the distance by 77 % — from a true $938$ pc to a spurious $1660$ pc.

Step 5. Finally, the reddening also fixes a gas column: $N_\mathrm{H}=5.8\times10^{21}\times E(B-V)=5.8\times10^{21}\times0.40=\mathbf{2.3\times10^{21}\ cm^{-2}}$ . True distance 938 pc, a 77 % distance error avoided, and $N_\mathrm{H}\approx2.3\times10^{21}\ \mathrm{cm^{-2}}$ of hydrogen inferred along the sightline — all from two magnitudes and a spectral type.

Physics of the portrait

Interstellar extinction is where the transparency of space stopped being assumed and started being measured. For most of the nineteenth and early twentieth centuries astronomers took the space between stars to be empty and clear; the person who proved otherwise, and in doing so rescaled the Galaxy, was an observer counting star clusters.

Origins · 1 minRobert Trumpler and how this was discovered

Robert Trumpler

1886–1956 · the discovery of interstellar absorption and reddening

Robert Julius Trumpler, a Swiss-born astronomer at Lick Observatory, settled a question that had lingered for a century: is interstellar space transparent? In 1930 he analysed a large sample of open (galactic) star clusters, for which he could estimate distances two independent ways — from the apparent brightness of their stars and from their apparent angular sizes. He found a systematic contradiction: the more distant a cluster, the larger it appeared than its brightness-based distance predicted. Rather than believe that distant clusters were physically bigger, Trumpler concluded that starlight was being dimmed by a pervasive absorbing medium concentrated toward the Galactic plane, so that brightness-based distances were being overestimated. He measured the effect at roughly $0.7$ magnitudes of general absorption per kiloparsec and, crucially, noted that the absorption was selective — it reddened the clusters, dimming their blue light more than their red. This single result did two things at once: it established the reality of interstellar dust, and it shrank the inferred scale of the Milky Way, because much of the apparent faintness of distant objects was absorption, not distance. The quantitative edifice of $A_\lambda$ , $E(B-V)$ , and $R_V$ that the rest of this article develops is the modern accounting of exactly the effect Trumpler forced astronomy to confront.

History

1930

— Trumpler proves interstellar absorption from the sizes of open clusters

1938

— Stebbins and Whitford map the reddening law by multicolour photometry

1965

— Stecher discovers the 2175 Å extinction bump from rocket ultraviolet spectra

1989

— Cardelli, Clayton and Mathis give the one-parameter R_V extinction law

1998

— Schlegel, Finkbeiner and Davis publish the all-sky reddening map

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., §12.1 (Interstellar Dust and Gas) · Binney & Tremaine, Galactic Dynamics, 2nd ed., §1.1 (dust and the appearance of the Galaxy) · Draine, Physics of the Interstellar and Intergalactic Medium, Chs. 21–23 (extinction and grain optics) · Ryden, Introduction to Cosmology, 2nd ed. (foreground extinction of standard candles) · Dodelson, Modern Cosmology (Galactic reddening in photometric surveys) · Mo, van den Bosch & White, Galaxy Formation and Evolution, §10 (the interstellar medium). The canonical numbers used throughout — $A_\lambda=1.086\,\tau_\lambda$ (with $1.086=2.5/\ln10$ ), $R_V\approx3.1$ for the diffuse ISM, and $N_\mathrm{H}/E(B-V)\approx5.8\times10^{21}\ \mathrm{cm^{-2}\,mag^{-1}}$ — rest on the CCM one-parameter form $A_\lambda/A_V=a(x)+b(x)/R_V$ (Cardelli, Clayton & Mathis 1989, ApJ 345, 245), the gas-to-reddening ratio of Bohlin, Savage & Drake (1978, ApJ 224, 132), and the limiting grain efficiencies $Q_\mathrm{ext}\to2$ (large grain) and $Q_\mathrm{ext}\propto\lambda^{-4}$ (Rayleigh limit), together with the distance bias factor $10^{A_V/5}$ . All-sky reddening maps are from Schlegel, Finkbeiner & Davis (1998), ApJ 500, 525.

Prerequisites. Interstellar Dust Grains · Phases of the Interstellar Medium · Magnitude and flux systems. Leads to. Thermal Dust Emission and Infrared Radiation · Interstellar Polarization and Magnetic Fields · Molecular Clouds and Giant Molecular Clouds · CO as a Molecular Gas Tracer and the X-Factor.

Ask the tutor: Ask me to deredden a real star from its photometry, or how a wrong $R_V$ biases a Type Ia supernova distance.

What is the difference between extinction and reddening, in one line each?

Extinction is the total dimming of a source by dust, measured in magnitudes as $A_\lambda$ at a given wavelength or as $A_V$ in the visual — it makes the star fainter. Reddening is the differential dimming, the fact that blue light is removed more than red, measured as the color excess $E(B-V)=A_B-A_V$ — it makes the star redder. They are two projections of the same wavelength-dependent extinction curve, and they are linked by the ratio $R_V=A_V/E(B-V)$ . You can measure reddening from colours alone, with no distance and no intrinsic-brightness knowledge, whereas measuring $A_V$ requires either $R_V$ or an independent brightness reference.

Why does the factor 1.086 keep appearing everywhere in extinction work?

It is simply the conversion between the natural-logarithm world of physics and the base-10 magnitude system of astronomy. Radiative transfer gives flux as $F=F_0 e^{-\tau}$ , an exponential in optical depth $\tau$ . Magnitudes are defined with $-2.5\log_{10}$ , so plugging the exponential in produces $A=2.5\,\tau\log_{10}e=2.5\,\tau/\ln10=1.086\,\tau$ . The number $1.086$ is exact: $2.5\div\ln10=2.5\div2.3026=1.0857$ . It means an optical depth of one corresponds to $1.086$ magnitudes of extinction, so to good approximation " $A_\lambda$ in magnitudes" and " $\tau_\lambda$ " are the same quantity in different units. Whenever you see it, the physics is just the log-base conversion.

Does interstellar reddening shift spectral lines like a cosmological redshift?

No, and conflating the two is a classic error. Reddening changes only the relative amount of flux at each wavelength — it tilts the continuum by removing more blue than red — but it leaves every spectral line exactly where it was, at its rest or Doppler-shifted wavelength. A redshift, by contrast, slides the entire spectrum, lines and continuum together, toward longer wavelengths by a factor $(1+z)$ . So a reddened star and a redshifted galaxy both "look redder", but they are physically opposite: one has had its blue light selectively eaten while its line positions stayed fixed, the other has had all its wavelengths stretched. You can always tell them apart by checking whether the line wavelengths moved.

Why is R_V about 3.1 in the diffuse ISM but larger in dense clouds?

$R_V$ measures the optical slope of the extinction curve, which in turn tracks the mean grain size. In the diffuse interstellar medium the grain-size distribution is dominated by small grains, giving a steep optical slope and $R_V\approx3.1$ . Inside dense, well-shielded molecular clouds the grains grow: they coagulate into larger aggregates and accrete icy mantles in the cold, dense gas. Larger grains scatter and absorb more "greyly" — with weaker wavelength dependence — which flattens the optical slope and pushes $R_V$ up to $4$ – $6$ . So a measured $R_V$ is a diagnostic of environment: a large value toward a cloud is direct evidence that its grains have grown, which is why $R_V=3.1$ should never be assumed blindly for cloud or cluster sightlines.

What causes the 2175 Å bump, and why does it matter?

The 2175 Å feature is a strong, broad absorption bump sitting on the ultraviolet extinction curve, and it is the most prominent spectral feature in Galactic extinction. It is attributed to $\pi\to\pi^*$ electronic transitions in small carbonaceous grains and polycyclic aromatic hydrocarbons (PAHs). It matters for two reasons. Diagnostically, its strength varies between environments — strong in the Milky Way, weaker in the Large Magellanic Cloud, and absent along some Small Magellanic Cloud sightlines — which probes the abundance and processing history of carbon dust in different galaxies. Practically, it lands right in the ultraviolet where many extragalactic and high-redshift observations are made, so getting the bump wrong biases dereddened UV fluxes and photometric redshifts. Its exact carrier remains an active research question.

How does dust extinction let us weigh molecular clouds?

Dust and gas are well mixed in the interstellar medium, so the dust column measured by extinction is a proxy for the total hydrogen column. The empirical relation $N_\mathrm{H}/E(B-V)\approx5.8\times10^{21}\ \mathrm{cm^{-2}\,mag^{-1}}$ , or equivalently $N_\mathrm{H}/A_V\approx1.9\times10^{21}\ \mathrm{cm^{-2}\,mag^{-1}}$ , converts a map of reddening directly into a map of gas column density. Observers build such maps by measuring the near-infrared color excess of many background stars shining through a cloud; each star's reddening gives the dust — and hence gas — column along its sightline, and the ensemble yields a spatial column-density map. Integrating that over the cloud's area and multiplying by the mean particle mass gives the cloud mass. This dust-based method is independent of, and complementary to, the CO-line technique, and it sidesteps CO's chemistry and excitation uncertainties.