Transcript Metallicity, Age, and Mass of Star Forming Galaxies at z~3

Dust and Its Effect on Galaxy SEDs
www.astro.umass.edu/~mauro/astro330/
What is Interstellar Dust? - 1
 Dust is formed from the most abundant elements in a galaxy, which
are also reactive: C, O, H (N is also abundant, but produces N2,
whose strong bond lowers reactivity)
 Si, Fe, Mg are major contributors, and provide most of the dust
mass together with C and O.
 Other elements also provide contributions (mostly minor, or trace).
 Gas condenses in dust; in the
MW the dust/gas~0.01;
 About 60% of C and 20% of O
is in dust; other elements are
shown in the Figure to the
right.
 Dust is roughly divided into
carbonaceous material (C and
H) and silicates (Si and O).
A.P. Jones, 1999
What is Interstellar Dust? – 2
 Dust forms in the ejecta and atmospheres of RGBs and
AGBs; possibly also in SNae (debated)
 Dust formation still a problem (models predict that too much
dust gets destroyed); this is complicated physics
 Most dust grains are small (a few hundred Å) and the
minority is `large’ (a few tenths of micron); the distribution of
grain sizes is roughly a power law: dn/da~a-3.5.
Polycyclic Aromatic
Hydrocarbons (C-H)
Weintgertner & Draine 2001
Large porous grains
Evidence for Interstellar Dust in
Galaxies
NGC4013
Cen A
NGC891
The Effects of Dust Grains on
Stellar Light
They absorb and scatter light.
Absorption heats the dust grains (~l-1.7 , which reflects the presence of a grain
size distribution), while scattering removes the light from the line of sight (~l-4).
Usually, the `pair method’ is used to determine the extinction
(=absorption+scattering)
Radiative Transfer of Dust - 101
ds
Incident light
In
Bn(T)
Scattered light
Q
Emerging (attenuated) light
Radiative Transfer of Dust - 101
Dust Extinction and Emission
Original Spectrum
Effect of dust
Energy absorbed by dust and re-emitted in the infrared
Case 1 – Single Point-Like Illuminating Source
(star, AGN, quasar, GRB, etc.)
At UV, optical, near-IR
wavelengths, the observer only
detects the light that comes along
the line of sight between the
observer and the source
(transmitted light).
Neither absorbed (IR) light, nor
scattered light is detected.
a)
The equation in the previous slide simplifies to:
Intrinsic, or non attenuated, spectrum
with solution:
Observed, attenuated spectrum
Attenuation factor
Expressing the optical depth t as a function of l :
where E(B-V) measures the depth (thickness) of the dust layer and
is the extinction curve
C. 2001
A Word About Extinction Curves
1. Systematic decrease from UV to IR;
it reflects dust grain size distribution.
2. Feature at 2175 Å is mainly
absorption (e.g., C. et al. 1995)
possibly due to PAHs (or graphite?)
3. Significant variations among
different lines of sight in the Milky
Way and between galaxies (MW,
LMC, SMC, M31; e.g., Cardelli,
SMC
LMC
small grains
MW
PAH? Graphite?
large grains
m = -0.4 log10 F(l)]
mobs (l) = mint (l) + Ao(l)
Clayton & Mathis 1989, Bianchi et al.
1996, Fitzpatrick 1999, Misselt et al.
1999)
Variations within MW usually leave the
2175 Å feature unchanged, thus possibly
due to variations in grain size distribution
(Cardelli et al. 1989, Draine et al. 2003).
Variations among galaxies often vary the 2175 Å feature.
Suggested to be destruction of carriers in active starforming regions (Gordon & Clayton 1998, Reach et al. 2000)
Examples of Spectra with Different
Reddening Values
Example of a hot
(blue) star,
extinguished by
increasing amounts
of Milky Way dust.
E(B-V)=0
E(B-V)=0.3
E(B-V)=0.6
The Effect of Dust in the Universe
Dole et al. 2006;
Hauser & Dwek 2001
n In
Dust absorbs light at
UV, optical, and nearIR wavelengths (with
decreasing crosssection) and re-emits it
in the mid/far infrared
(>3-4 mm)
Stellar light
Dust-processed light
About equal quantities of luminous energy in the Universe (integrated along
the line of sight over all redshifts) emerge as direct stellar light and as dustprocessed light.
And A Word About Common Methods to
Measure Dust Reddening
If the star (or stellar population) is sufficiently young and massive to produce
ionizing photons, a common method is to measure the deviation of hydrogen
recombination line ratios from theoretical expectations.
These have well established values, that vary typically less than 10%-15% for a large
range of physical (temperature, density) and chemical conditions.
For instance, the two strongest lines at optical wavelengths are Ha (6563 Å) and Hb
(4861 Å). Their intrinsic luminosity ratio:
which can be compared with the observed ratio, to derive the optical thickness of the
dust layer, E(B-V):
Multiple and well-separated (in l) lines generally yield the most accurate measures.
Hydrogen Atom Energy Levels
Some Examples of Emission Line Pairs
C. 2001
When measures for at least 3 lines are available, especially widely separated in
wavelength (e.g., Ha, Pab, Brg), the geometry of the dust distribution can be
constrained.
This becomes important for complex source-dust distributions (see next few slides),
when different lines probe different depths in the dust.
Case 2 – Complex, Extended Illuminating
Source (galaxy, star forming region, etc.)
b)
a. Reddening/attenuation of light
because of :
I. Extinction
b. `Bluing’ of light because of:
I. Scattering into line-of-sight
II. Dust clumpiness
III. Differential optical depth
In complex systems, UV/optical dust reddening is unrelated to dust attenuation!!
A highly influential work by Witt, Thronson & Capuano (1992) was one of the first to
quantify the impact on the observed colors of the complex dust distribution in galaxies.
Quoting:
`One explanation for the widespread adoption of a mythological overlying screen of dust
absorption in galaxies is a parallel with stellar astronomy, in which only grains along the
line of sight to a star will extinguish the light.’
A Simple Example: Dust and Stars
Homogeneously Mixed
What is the observer going to see?
Homogeneous mixture
E(B-V)
Foreground screen
Let’s Interpret the Previous Results in
Terms of Radiative Transfer
For the foreground screen, we can use the same equation and solution as for single point
sources:
with: t = 0.921 E(B-V) kn
and Ion = input intensity (in the absence of dust)
For the homogeneous mixture of dust and stars, we need to introduce a source function in
the radiative transfer equation:
with:
In this case, the optical depth t is an `effective’ optical depth, that needs to include the
effects of scattering into the line of sight (Mathis 1983).
Dust Geometry is NOT the Only Problem:
AGE-DUST Degeneracy
Compare the red and black
models:
A young, dusty stellar population
can have a UV-optical-nearIR
Spectral Energy Distribution
similar to that of an old, dust-free
stellar population.
The only obvious discriminants,
which may or may not be
observable, are the ionized
emission lines and the 4,000 Å
break.
Plus, of course, the mid/far IR
emission!
Starburst Attenuation
High
Attenuation
ATTENUATION
REDDENING
A high value of IR/FUVobs indicates a
large fraction of UV light reprocessed in
the IR.
Low
Attenuation
BLUE
Meurer et al. 1999
RED
~ b26
Termed the IRX-b relation
The plot to the left shows that for UVbright starbursts, high attenuation is
accompanied by high reddening (like
single sources?)
Starburst Attenuation Curve
Obscuration, attenuation, …. all
attempt to describe the loss of
UV/optical light to the IR.
The same curve can account for
the observed reddening as well.
The above means that both the
absolute values and relative
values of the curve can be used
to describe the dust impact in
the UV-bright starbursts.
C. et al. 1994, 1996, 2000; C. 1997, 2001
The absence of the 2175 Å is
real, possibly an effect of the
destruction of the carriers in the
starburst environment.
Implementation of the Starburst
Attenuation Curve
R’V = 4.05
Used as:
fobs(l) = fint(l) 10{-0.4 k’(l) [0.44 E(B-V)]}
in the models that we will use to fit the spectrum of NGC7714