1 Astro 6590: Galaxies and the Universe Lecture #2 (Wed Sep 3, 2008) • Extinction and reddening • Overview of stars • Star counts • Luminosity functions • Malmquist bias • Overview of the H-R diagram • Color magnitude diagrams (CMDs) • Overview of stellar evolution in the CMD • Spectroscopic parallax • Overview of stellar spectra • Overview of galaxy spectra • The 4000 Å break • Initial mass function • Star formation rate • Overview of population synthesis • Case studies: Palomar 11, Sex A, LGS3 Photometric observations What we observe is: • extincted/fainter • absorbed/reddened • redshifted • projected • convolved • evolved m obs = m intrinsic + ∆m ext + ∆m z a) Extinction by the Earth’s atmosphere (airmass) b) Extinction within the Milky Way (Galactic extinction) c) Extinction within external galaxy (internal extinction) d) Redshift due to Hubble expansion (“K correction”) Extinction Atmospheric extinction Interstellar extinction Let α λ be the scattering cross section at wavelength λ; the optical depth τ λ at distance d is: τ λ = ∫ α λ n(s) ds Observed flux f obs,λ relative to the emitted f 0,λ => f obs,λ = f 0,λ e -τ Since = 10 -0.4 = e - => star dimmed by ∆m = 1.086 τ λ f 0,λ f obs,λ (m 0 -m obs ) τ λ d 0 The importance of dust • Star formation : stars form from dusty molecular clouds • Dust scatters and absorbs light • Effective absorber of photons of λ ≤ a (a = size of grain) • Most grains have a > ~ 0.1 μ => UV more effected “classical” dust grains: a ~ 0.1 μ and N > 10 4 atoms “small” dust grains: N < 100 atoms PAH’s (Polycyclic Aromatic Hydrocarbons) N ~ 50 atoms Leger and Puget (1984): Large planar molecules • Galactic Center: A V ~ 30 mag => 1 photon in 10 12 penetrates A 2.2μ ~ 2.5 mag => 1 photon in 10 penetrates • Dust polarizes light • Observed starlight is linearly polarized due to population of paramagnetic dust grains partially aligned with IS B-field. Spectra fit by modified Planck curves I ν ∝ε ν B ν (T) and ε ν ∝ν n Wien’s law: λ peak = 250 μ (20 K/T) 5 μ 1000K 50 μ 100K 250 μ 20K λ peak ~ 1mm T = 5K
21
Embed
a590 08 2hosting.astro.cornell.edu/.../astro590/pdf08/a590_08_2.pdf · 2008. 9. 2. · d) Redshift due to Hubble expansion (“K correction”) Extinction Atmospheric extinction Interstellar
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Astro 6590: Galaxies and the UniverseLecture #2 (Wed Sep 3, 2008)
• Extinction and reddening• Overview of stars• Star counts• Luminosity functions• Malmquist bias• Overview of the H-R diagram• Color magnitude diagrams (CMDs)• Overview of stellar evolution in the CMD• Spectroscopic parallax• Overview of stellar spectra• Overview of galaxy spectra• The 4000 Å break• Initial mass function• Star formation rate• Overview of population synthesis• Case studies: Palomar 11, Sex A, LGS3
mobs = mintrinsic + ∆mext + ∆mza) Extinction by the Earth’s atmosphere (airmass)b) Extinction within the Milky Way (Galactic extinction)c) Extinction within external galaxy (internal extinction)d) Redshift due to Hubble expansion (“K correction”)
ExtinctionAtmospheric extinction
Interstellar extinctionLet αλ be the scattering cross section at wavelength λ; the optical depth τλ at distance d is:
τλ = ∫ αλ n(s) dsObserved flux fobs,λ relative to the emitted f0,λ => fobs,λ = f0,λ e-τ
Since = 10-0.4 = e- => star dimmed by∆m = 1.086 τλ
f0,λfobs,λ
(m0-mobs) τλ
d0
The importance of dust• Star formation: stars form from dusty molecular clouds• Dust scatters and absorbs light
• Effective absorber of photons of λ ≤ a (a = size of grain)• Most grains have a > ~ 0.1 μ => UV more effected
“classical” dust grains: a ~ 0.1 μ and N > 104 atoms“small” dust grains: N < 100 atomsPAH’s (Polycyclic Aromatic Hydrocarbons) N ~ 50 atoms
Leger and Puget (1984): Large planar molecules• Galactic Center: AV ~ 30 mag => 1 photon in 1012 penetrates
Galactic extinction law: determination of Rλ•Consider two *s of the same spectral type (same intrinsic SED), one nearby, unreddened by extinction and the other further away and reddened. Observe their mags at several λs to get color excesses: e.g., E(U-V), E(B-V), E(R-V), E(I-R) etc.
Color excesses will depend on amount of reddening of reddened *•Normalize to unity E(B-V)
and plot E(λ-V) as f(λ):
•It has been shown that along •l.o.s. not crossing dense clouds:
RV ≈ 3.2 ± 0.02
Johnson 1965
Level of reddening curve extrapolated to 0 absorption at λ = ∞
E(X
-V)
AV
3
Extinction lawsExtinction law = A(λ)/A(V)
Parameterization: the average Rv-dependent Extinction Law
A(λ)/A(V) = a(x) + b(x)/Rv(x= λ-1)
Rv affects the shape of the extinction curves(particularly at the shorter wavelengths)
Cardelli, Clayton & Mathis (1989, ApJ 345, 245) show plots of Aλ/AV as a f(RV), where
(a) as λ → UV, the ratio Aλ/AV ↑(b) at fixed λ in the UV, Aλ/AV ↑ as RV ↓
The differences among different lines of site become more dramatic as λ → UV
Cardelli et al’’s extinction law
Cardelli et al. 1989; Mathis 1990 ARAA
A bump around2175 Å => graphite resonance?
Serious deviationfor x > 7 μm-1
Shapeindependent ofRv in the NIR
U.B.V bandsfor reference
Shown for 3 different l.o.s.
Major IR instruments/missions
IRAS (InfraRed Astronomical Satellite) : Jan 1983 launch• 90% of sky at 12, 25, 60 and 100 μ• ULIRGs: emit most of their light in FIR• IR “cirrus”: diffuse, high latitude, 100 μ• Often use 2 component model
1. T ~ 170K, n=1: hot dust, associated with SF2. T ~ 30K, n=2: cool dust, associated with cirrus
COBE (Cosmic Background Explorer): Nov 1989 launch• FIRAS: precision measurement of CMB spectrum• DMR: search for CMB anisotropies on 7º scale• DIRBE: cosmic IR background (absolute brightness) in 10
bands from 1.25 to 240 μNIR bands (1-5 μ): starlightMIR bands (12, 25, 60 μ): interplanetary dust, warm IS dust (T ~ 40-60K)FIR bands (100, 140, 240 μ): Galactic, cool IS dust
100 μ: T ~ 20-25K; 140,240μ: T ~ few tens of K.
Corrections to observed magnitudes
using the DIRBE instrument on COBE
4
Star counts, LFs etc Galactic longitude and latitude
Binney & Tremaine
Star counts in the Milky Wayν(r,l,b,M,S): space density of stars of absolute magnitude M and spectral type S at distance r in the direction l,b. A(m,l,b,S): star counts of stars of spectral type S in direction l,bAlong any l.o.s. (l,b) then, assume that:
ν(d,M,S) dM dV = Φ(M,S) dM DS(r) dV
where dV is a volume element, dM is an increment of absolute magnitude, DS(r) is the relative density function (density of stars of spectral type S at distance r in units of the corresponding stellar density in the solar nbhd) and Φ(M,S) is the luminosity function (actual number of *s of M and S per cubic pc in the solar neighborhood).
Most methods for determining L.F. involve counting the number dN/dm of objects that have m in the range (m, m+dm) and that lie in some area of the sky. The star count function then is
A(m) ≡
What do we expect the star counts to show?
dNdm
Star countsA(m,S): star counts of stars of spectral type S in a direction l,bIf some * has M, it will appear to have m if it is a distance r:
M = 5 log r + a(r) + M – 5where a(r) is the absorption in magnitudes along the l.o.s.
In a field subtending ω, stars in the distance interval (r, r+dr) occupy a volume dV = ωr2dr. Hence from this volume, we obtain a contribution to the star counts at apparent magnitude m of
dA(m,S) = Φ(m+5-5log r – a(r),S) DS(r) ωr2drStars at another m’ and r’ will make a similar contribution.Thus, the total # of *s A(m,S) is obtained by summing over all distances or, equivalently, over stars of all M distribute along the l.o.s. in such a way that m = const.
A(m,S) = ω ∫ Φ[m+5-5log r-a(r),S] DS(r) r2drFor general star counts, lumping all spectral types, we get
A(m) = ω ∫ Φ[m+5-5log r-a(r)] D(r) r2dr
Mihalas & Binney 4-2
∞0
∞0
5
Stellar luminosity functionThe general luminosity function then is
Φ(M) ≡ Σ Φ(M,S) and D(r) is the total # of *s of all types per unit V at distance r, inunits of the total # of all types per unit V in the solar nbhd.
=> This is true only if the LF is constant along the l.o.s. (same as in solar nbhd).
Mihalas & Binney 4-2
What do you know about the Hipparcossatellite?
Galaxy luminosity functions
Solid line shows the Schechter fit for L* ~ 8x109h-2 L⊙
Luminosity functions
• Elliptical Gals brightest• Sa’s brighter than most Sc’s
Malmquist bias
See BM 3.6.1 for derivation (and basis for homework problems)
6
Malmquist bias: practical implications
Roberts & Haynes 1994, ARAA 32, 115
What do you notice?
How do you explain it?
Malmquist bias: practical implications
HIPASS Completeness Limit
HIPASS Limit
HIPASS bandwidth
RG, MH & ALFALFA collaboration
Selection function
Example: Suppose you want to investigate the galaxy density in some some sample which has a limiting magnitude mlim (or flim).Nearby, you will detect all of the galaxies brighter than Mlim (or Llim), but further away, you will ONLY detect the bright galaxies (M<Mlim (or L>Llim).
.
ϕ(x): “selection function”: an estimate of the probability that a galaxy brighter than Mlimat distance x>R is included in the sample:
ϕ(x) = ∫ Φ(M)dM
∫ Φ(M)dM
M(x)
Mlim
-∞
-∞
e.g., Huchra & Geller 1982, ApJ 257, 423
ϕ(x<R) =1
Hipparcos CMDHipparcos satellite
Along the MS:
L ∝ M3.5
What do you know about this satellite?
7
Temperatures of stars Spectral class, color and Teff
M5M0K5K0G5G0F5
F0A5A0B5B0O5
Spectral Class
+10.0+9.0+7.3+5.9+5.1+4.4+3.4
+2.6+2.0+0.7-1.1-4.1-5.7
M.S. MV
+1.63+1.45+1.18+0.89+0.70+0.58+0.42
+0.27+0.130.00-0.16-0.31-0.35B – V
Reln between (B-V) and Teff from Fowler 1996, ApJ
Different types offset 0.3 dex
Color-magnitude diagrams (CMDs)The horizontal axis of the H-R diagram can be represented in different ways:
• Spectral class• Effective surface temperature• “Color index”: difference in brightness measured in
separate (wavelength) filters
8
Basic relations Stellar spectral types
Stellar luminosity classes
Why?
The method of “spectroscopic parallax”
1. Observe the star’s apparent brightness.
2. Observe the star’s spectrum; determine its spectral class and luminosity class.
3. Place the star on the H-R diagram; estimate its luminosity.
4. Use luminosity and apparent brightness to get distance.
Also works for star clusters, galaxies
9
Color-magnitude diagrams (CMDs)
•Also useful as a distance indicator • Extragalactic distance scale discussion (later)
Evolution of High Mass Stars• A high mass star also shines by burning hydrogen when it is on
the Main Sequence, but• The sequence of fusion reactions (still 4H => He) is
different (the “C-N-O cycle” not the proton-proton chain).• Although a high mass star has more hydrogen to begin with, it
burns its hydrogen must faster than a low mass star does.• High mass stars have shorter Main Sequence lifetimes than
low mass stars.
• The basic sequence of Post Main Sequence Evolution is the same for high and low mass stars, until the star reaches the Red Supergiant Phase.
• In a low mass star, the carbon-oxygen core cannot burn further.• In a high mass star, the core temperature and density are so
high that further burning sequences take place.
Evolutionary tracks in the H-R diagram
Girardi et al 2000, AApSuppl 141, 371
• These are tracks only for low mass stars (mass indicated at initial point of post-MS evolution)
• The timescale for the complete track is 25 Gyr.
• Only the high mass stars are really interesting!
10
Evolution depends on metal abundance
Girardi et al 2000, AApSuppl 141, 371
• Age range goes from log (t/yr) = 7.8 to 10.2 at equally spaced intervals of ∆log (t/yr) of 0.3
• Left: metal poor population; Right: solar abundance
X+Y+Z=1
Galaxy colors and stellar evolution• Stars have masses in range ~0.07 to 50 M• Main Sequence: stable H-burning phase
- 2.7 for 1.0 ≤ M < 120- 2.3 for 0.5 ≤ M < 1.0- 1.3 for 0.08 ≤ M < 0.5- 0.3 for 0.01 ≤ M < 0.08
x =
# of *s w/ masses in (M, M+dM) formed at a given time in a given volume
Star formation rate (SFR)Suppose start with initial mass of gas Mg,0
SFR => R(t,τ) = rate at which the mass of gas is convertedinto stars per unit time
Constraint: ∫ R(t,τ) dt = Mg,0
Star formation models:• Constant SFR until some time t = τ, then halt
R(t,τ) = 0 < t ≤ τ
R(t,τ) = 0 t ≥ τ
• Exponential
R(t,τ) = exp (-t/τ)
∞
0
Mg,0
τ
Mg,0
τ
If τ is short:“burst”
Stellar Mass from Colors
Bell et al. 2005, ApJ 625, 23
Sextans ADwarf Irregular
D=1.44 Mpc
17
CMD of Sextans A
Actively forming stars for last 2.5 Gyr; increased rate < 100 Myr ago. Some old stars (but images not deep enough to track details)SF found in 3 zones (high NHI); youngest is single region < 20 Myr old
Also Dolphin et al. 2003, AJ 126, 187
CMD of Sextans A
Also Dolphin et al. 2003, AJ 126, 187
LGS 3Miller et al 2001, ApJ 562, 713
• Stellar population dominated by old, metal poor stars but….• Handful of young blue stars + HI gas• Transition object between dSph and dI?
D = 760 kpc
Small amount of HI (2 x 105 M⊙)
LGS 3Miller et al 2001, ApJ 562, 713
• Stellar population dominated by old, metal poor stars but….• Handful of young blue stars + HI gas• Transition object between dSph and dI?
Stellar population and star formation history vary with radial position in the galaxy
Most of the young stars are in central 63 pc where the SFR has been fairly constant. Elsewhere: declining SFR
18
LGS 3 versus Sex A Galaxy spectra
• Redshift
• Velocity dispersion/rotational velocity
• Star formation rate
• AGN activity
• Abundances
Observing galaxy spectra
The observed spectrum of a galaxy, made up of a large # of *s, (assumed here, for simplicity, to be identical) can be described as
G(u) = ∫ dvlos F(vlos) · S(u – vlos)where u = c ln λ is the wavelength expressed in logarithmic units
S is the spectrum of the star in the same unitsF(vlos) describes the distribution of the stellar line-of-sight
velocities within the portion of the galaxy observed.
Note: this is a convolution integral. In practice, it is therefore possible to extract F(vlos) from an observed G(u) if a suitable template stellar spectrum S is available.
Basic assumptions:• Galaxies can be treated as closed systems• Chemical enrichment is not important after stars form• The SFR is either a burst or a smooth function of time,
independent of stellar mass; it determines the spectral and luminosity evolution of the galaxy
• The IMF is a simple function of the stellar mass (independent of galaxy age)
• The effect of dust and gas on the observed spectra can be treated separately
We will look at B&C again later
Models vs. Data: Spirals
• Increasing SF time scales produce good fits for increasingly late-type Spirals
• Note good fit of 4000 Å break and main absorption features in all cases
BC93 Fig. 7Bruzual & Charlot 1993, ApJ 405, 538
Models vs. Data: Ellipticals
• Best-fitting age model and composite elliptical spectrum
• Fairly good fit over entire spectral range
• Note UV-rising branch, highlighting importance of accurate AGB modeling
• ELLIPTICALS: The oldest objects formed stars within 1-2 Gyr of the BB and have had little SF since. “RED AND DEAD”