26 February 2013 Astronomy 142, Spring 2013 1

Today in Astronomy 142: interstellar dust

Extinction, reddening and polarization of starlight Composition and structure of dust grains

The sky, dominated by the Milky Way (Axel Mellinger, Central Michigan U.)

26 February 2013 Astronomy 142, Spring 2013 2

How do we know that there is interstellar dust?

Extinction: dark markings on the Milky Way (dark clouds) are generally absorption and/or scattering of background starlight, rather than holes in the distribution of stars. Reddening: stars associated with dark markings are often much redder in color (in their continuum) than would be inferred from their spectral type (classified by their line spectra). This selective extinction is naturally explained by the wavelength dependence of scattering and absorption by submicron-size solid particles. Polarization of (originally unpolarized) starlight: naturally explained by selective absorption by nonspherical, aligned, submicron-size solid particles.

Dark clouds

Like much else in astronomy, dark clouds were “first” seen by William Herschel. According to his sister and collaborator Caroline, he first saw these, in Scorpio and Sagittarius (1774), saying to her, “hier ist wahrhaftig ein Loch im Himmel” (surely here is a hole in the sky). 26 February 2013 Astronomy 142, Spring 2013 3

Stéphane Guisard, ESO

Antares (α Sco)

Trifid Nebula (M 20)

26 February 2013 Astronomy 142, Spring 2013 4

Dark clouds

(continued)

Visible-light photograph by David Malin (AAO)

Good, modern images reveal filamentary, cloudlike structure to the dark markings, which certainly makes it seem more plausible that they are dark, absorbing clouds rather than holes in the distribution of stars.

26 February 2013 Astronomy 142, Spring 2013 5

Dark clouds (continued)

Visible-light image by Jean-Charles Cuillandre (CFHT).

And there are many examples of “holes” in star fields or nebulosity so dense that it would be implausible for them to exist as holes long enough for us to see as many as we do.

NGC 6520

Barnard 86

Dark clouds (continued)

26 February 2013 Astronomy 142, Spring 2013 6

α Cen

β Cen

Crux

The Coalsack

Yuri Beletsky (ESO)

26 February 2013 Astronomy 142, Spring 2013 7

Dark clouds (continued)

NOAO

The Horsehead Es ist nicht ein Loch im Himmel.

NGC 2024

Orion’s belt

Aber hier ist wahrhaftig ein Loch im Himmel.

26 February 2013 Astronomy 142, Spring 2013 8

Here is an exception which proves the rule: NGC 1999 in Orion. Long thought to be a

dark globule overlying a more diffuse cloud, and extinguishing starlight scattered from the latter.

The star would lie further from us than the globule, nearer than the diffuse cloud. Visible light, HST/STScI/NASA

Hier ist wahrhaftig ein Loch im Himmel (continued)

But the “globule” turns out to be dark even at very long infrared and submillimeter wavelengths. An ordinary globule

would emit brightly at such long wavelengths.

So in this case the thing that looks like a hole really is a hole.

Stanke et al. 2010. Blue = 4.5 µm (Spitzer Space Telescope); green = 70 µm, red = 160 µm (Herschel Space Observatory). 26 February 2013 Astronomy 142, Spring 2013 9

HST

26 February 2013 Astronomy 142, Spring 2013 10

Proof of the existence of extinction: Trumpler’s experiment on open clusters

Plot sum of fluxes from all stars in cluster vs. the solid angle the cluster occupies. If all open clusters had the same diameter D, then , and since flux also scales with 1/r2 , he should get a straight line through the origin. He got:

( )Ω = ∝π D r r2 12 2

Total flux

Scatter: they don’t all have the same D.

Curve doesn’t pass through origin: flux falls faster than 1/r2 => extinction.

Solid angle

( )∝ 1 2r

( )∝ 1 2r(Trumpler 1930)

26 February 2013 Astronomy 142, Spring 2013 11

Rayleigh scattering

Results for scattering by non-absorbing dielectric spheres with refractive index n and radius a λ, on the flux from a background source that would give a flux f0 in the absence of dust:

( )0 0

23

4

, where

32 11 mean free path

3 optical depth in scattering

sx

s

f f e f e

n

Nx

τα

πα

λτ α

−−= =

−= =

= =

*

x

f N particles/unit volume

Star

Dust cloud

26 February 2013 Astronomy 142, Spring 2013 12

Rayleigh scattering (continued)

These expressions will be derived for you in PHY 218. We won’t use them further but they serve as a useful illustration of the strong wavelength dependence of light scattering by small particles. Note that the larger τs is, the less light is transmitted, and the more light is scattered into other directions. Also note that short-wavelength light is scattered more effectively than long-wavelength light, because . Thus the sky is blue, as are reflection nebulae.

By the same token, what gets transmitted is redder, because the blue is scattered away from the light’s original direction. Thus sunsets are red, as is other extinguished starlight. Longer-wavelength light (e.g. infrared, radio) can see

through dust.

41sτ λ∝

26 February 2013 Astronomy 142, Spring 2013 13

Trifid Nebula (M20)

Photograph by David Malin (AAO). Note: Dark lanes in upper,

ionized, nebula (extinction)

Blue color of lower, reflection, nebula (scattering)

Many bright red stars in field (reddening)

26 February 2013 Astronomy 142, Spring 2013 14

Polarization of light scattered by nonspherical dust grains

Interstellar dust grains are usually far from spherical: they tend instead to be needle- or flake-like. Thus they can absorb or scatter light with some polarizations – the components of E along the long dimension of the grain – better than others.

Light

Dust grain seen from side

Dust grain seen from top

Scattered light polarized along grain direction

Transmitted light polarized perpendicular to grain direction

26 February 2013 Astronomy 142, Spring 2013 15

Polarization of light scattered by nonspherical dust grains (continued)

Interstellar dust grains are often aligned with their long axes along some given direction. That direction can be determined by external magnetic fields and/or gas motions. Most common alignment: B perpendicular to the long axis of spinning

dust grain (Davis & Greenstein 1951). Below: electric polarization of stars in different distance ranges, as a function of Galactic longitude and latitude (Axon and Ellis 1976). The orientation shows that B is mostly parallel to the plane of the Galaxy.

26 February 2013 Astronomy 142, Spring 2013 16

Extinction and reddening by real dust grains

Most interstellar grains aren’t just dielectric; they absorb light, too. Empirical relation for τa: (except for certain special wavelengths - see below). Reddening – or differential extinction – is defined by the color excesses, and where

1.85 , 0.5 20 maτ λ λ µ−∝ = −

0.5 0 0.5 1 1.5 21.5

1

0.5

0

0.5

1

1.5

B-VU

-B

Reddening correction

EB V−

( )( )

0.72 .E U BE B V

−≅

−

U BE −

Empirical color-color relation for unextinguished zero-age main sequence stars

( )E U B−( ) ,E B V−

26 February 2013 Astronomy 142, Spring 2013 17

Reddening correction

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0-0.5 0.0 0.5 1.0 1.5 2.0

B-VU

-B

Nearby main-sequencestarsCluster withcolorcorrection

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0-0.5 0.0 0.5 1.0 1.5 2.0

B-V

U-B

Nearby main-sequencestarsDistant opencluster

Take observations at three wavelengths (e.g. U, B, V); compare the color-color plot to unextinguished stars.

Shift the plots until they fit; the amounts by which the cluster shifts are E(B – V) and E(U – B): values 0.32 and 0.18, here.

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0-0.5 0.0 0.5 1.0 1.5 2.0

B-VU

-B

26 February 2013 Astronomy 142, Spring 2013 18

Color excess and extinction

The B-V color excess is related to extinction optical depth. Empirically, for diffuse-cloud extinction: In magnitudes: where, for diffuse dark clouds, R = 3.06. Effects: Stars look too red for their temperature. Stars look too faint for their distance.

2.76V B VEτ −=

( )

( ) ( )

( )

0

0/5

0 02.5 100V V

V

V V V

A AV V V

A RE B Vm m AB V B V E B V

f f f− −

= −

= −

− = − − −

= = ×

26 February 2013 Astronomy 142, Spring 2013 19

Extinction correction

Recipe: Reduce every B-V by the

same amount, . Reduce every V by the

visual extinction . The whole HR diagram

shifts to bluer and brighter values.

You’ll be doing some of this in Homework 5, this week’s recitation, and future homework sets.

8

10

12

14

16

18

20-0.5 0 0.5 1 1.5

B-VV

( )E B V−VA

B VE −

VA

26 February 2013 Astronomy 142, Spring 2013 20

Interstellar dust grains themselves

Size: more like smoke than household dust. < 0.1 µm in most dark clouds, up to ~1 µm in the darkest molecular clouds, and down to 0.001 µm (50-100 atoms) in diffuse clouds and UV-illuminated clouds. Amount: about 1% by mass of the interstellar medium (see next class). Made of: silicates, carbon and probably metallic iron, just like terrestrial planets. Temperature: T = 10-100 K. Heated by ultraviolet starlight. Cooled by blackbody emission. The radiation by inter-

stellar dust grains can be seen at infrared wavelengths.

Interstellar silicate dust grains

Especially evident at wavelengths 9.7 and 18 µm, interstellar silicates are submicron in size, and amorphous (not crystals). By mass, > 96% magnesium silicates, half Mg2SiO4, half

MgSiO3; ~3% SiC, < 1% crystalline silicates. Highly irregular in shape. (Min et al. 2007)

26 February 2013 Astronomy 142, Spring 2013 21

GCS 3-I; Chiar & Tielens 2006. Emission lines at 12.8, 18.7, 33.4 and 34.5 µm are not related to the dust.

5 20 2 50

9.7 µm 18 µm

Interstellar carbon dust grains

Also submicron, and extinguishes due both to amorphous and molecular components, both pure and hydrogenated. Most distinctive feature of the

bulk of carbon grains is a 217.5 nm absorption due to graphite.

Largest contribution to extinction is due to continuous absorption: both graphite and amorphous carbon are electrical conductors.

26 February 2013 Astronomy 142, Spring 2013 22

Massa and Fitzpatrick 1986

26 February 2013 Astronomy 142, Spring 2013 23

Interstellar carbon dust grains (continued)

The smallest carbon dust grains only have tens of atoms, and have several unusual aspects: Only seen by emission features,

never in absorption. Appear hot even when far from a

star, because that heating requires only single UV photons (Sellgren 2004).

Pattern of infrared features indicates that these grains are polycyclic aromatic hydrocarbons (PAHs; Leger and Puget 1984).

A typical PAH: coronene, C24H12

Werner et al. 2004

Origin and life of interstellar dust grains

Grains are born primarily in the winds of late type (giant, AGB) stars, condensing as the wind material cools, and perhaps secondarily in the interstellar medium, condensed from cold gas and perhaps conglomerated with other grains. Dust often seen to be crystalline in giant stars, yet it is not

crystalline in the interstellar medium. But the average dust grain lives billions of years in the

interstellar medium, enough time for originally crystalline grains to become shattered and amorphotized by UV photons and cosmic rays.

Cold-gas condensation and conglomeration naturally produce amorphous grains.

(See, e.g. Draine 2006, 2009) 26 February 2013 Astronomy 142, Spring 2013 24