Transcript Interstellar Medium and Star Formation

Interstellar Medium and Star
Formation
Astronomy G9001
Prof. Mordecai-Mark Mac Low
Historical Overview of
Observations



Dust
Excess Mass
Visual Nebulae
– Emission lines
– Continuum light


Polarization
Optical Absorption
Lines






HI lines, & radio
continuum
UV Absorption lines
X-ray emission
Molecular line
emission
IR emission
Gamma Rays
Following Li &
Greenberg 2002,
astro-ph/0204392
Dust
Naked eye observations of dust clouds
 Holes in the heavens (Herschel 1785) vs
obscuring bodies (Ranyard 1894, Barnard
1919)

– Partial obscuration of continuous nebulae
– Smooth dimming of star fields
– Shapley-Curtis debate 1920
• Shapley saw no obscuration in globulars: but they
were out of plane!
• Does obscuration contribute to distance scale?
Reddening
Extinction was known since 1847 (though
not taken seriously in Galaxy models)
 Reddening discovered by Trumpler (1930)
 Wavelength dependence established
obscuration as due to small particles
 Reddening proportional to NH

– Extremely high NH measurable in IR against
background star field: NICE (Lada et al. 1994,
Cambrésy et al. 2002).
Excess Mass
Vertical stellar motions allow
measurement of non-stellar disk mass
 Excess density of 6 x 10-24 g cm-3
found by Oort (1932)
 We now know that this is a
combination of ISM and dark matter.
 Similar methods still used to measure
dark matter density.

Visual Nebulae




Nebulae first thought to be stellar
Spectroscopy revealed emission lines from
planetary nebulae, establishing their gaseous
nature (Huggins 1864)
Reflection nebulae distinguished from emission
nebulae by continuous spectrum, reddening of
internal stars
Measurements of Doppler shifts in emission
lines revealed supersonic turbulent motions in
Orion emission nebula (von Weizsäcker 1951, von
Hoerner 1955, Münch 1958).
Polarization
General linear polarization of starlight by ISM
discovered by Hill (1949) and Hiltner (1949).
 Alignment of dust in magnetic field (tho
mechanism remains debated)
 Revealed large scale field of galaxy
 Radio polarization of synchrotron shows field
in external galaxies as well
 At high extinctions (high densities), IR
emission polarization fails to trace field

(Goodman et al. 1995)
Optical Absorption Lines

Ca II H & K lines have different dynamics
from stellar lines in binaries (Hartmann 1904)
– Na I D lines behave similarly (Heger)
– Now used to trace extent of warm neutral gas
– Reveals extent of local bubble (Frisch & York
1983, Paresce 1984, Sfeir et al 99)

Lines spread over 10 km/s, although
individual components only 1-2 km/s wide
– Interpreted as clouds in relative motion
– Reinterpretation in terms of continuous
turbulence?
HI lines
HI fine structure line at 21 cm (Ewen &
Purcell 1951) reveals cold neutral gas (300 K)
 Pressure balance requires 104 K intercloud
medium (Field, Goldsmith, Habing 1969)
 Large scale surveys show

– Supershells and “worms” (Heiles 1984)
– Vertical distribution of neutral gas (Lockman,
Hobbes, & Shull 1986)

Distribution of column densities shows
power-law spectrum suggestive of
turbulence (Green 1993)
Radio Continuum
First detected by Reber (1940): Nonthermal
 Explanation as synchrotron radiation by

Ginzburg
Distinction between thermal (HII regions)
and non-thermal (relativistic pcles in B)
 Traces ionized gas throughout Milky Way
 Evidence for B fields and cosmic rays in
external galaxies

UV Absorption Lines
Copernicus finds OVI interstellar absorption
lines (1032,1038 Å) towards hot stars
 Photoionization unimportant in FUV
 Collisional ionization from 105 K gas, but
this gas cools quickly, so must be in an
interface to hotter gas
 First evidence for 106 K gas in ISM

X-ray emission
Confirms presence of hot gas in ISM
 Diffuse soft X-ray background (1/4 keV)
anticorrelates with NHI: Local Bubble
(McCammon et al. 1983, Snowden et al. 1990)
 Detection of SNRs, superbubbles
 X-ray shadows of cold clouds show
contribution from hot halo (Burrows &

Mendenhall 1991, Snowden et al. 1991)
Molecular line emission
Substantial additional mass discovered with
detection of molecular lines from dense gas
 Millimeter wavelengths for rotational, vibrational
lines from heterogeneous molecules
 NH2 and H2O first found (Cheung et al. 1968,
Knowles et al. 1969) then CO (Penzias et al. 1970),
used to trace H2
 Superthermal linewidths revealed (Zuckerman &
Palmer 1974) showing hypersonic random motions
 Map of Galactic CO from roof of Pupin (Thaddeus

& Dame 1985)
IR emission
Only with satellite telescopes such as IRAS
was IR emission from cold dust in the ISM
detectable: the “infrared cirrus”
 IR penetrates dust better than visible, so it
allows observation of star formation in
dense regions

Gamma Rays

Gamma ray emission from Galactic plane first
detected with OSO 3 and with a balloon (Kraushaar et
al. 1972, Fichtel et al. 1972)


Confirmed by SAS 2 and COS B at 70 Mev.
CR interactions with gas and photons:
– Electron bremsstrahlung
– Inverse Compton scattering
– Pion production

Independent estimate of mass in molecular clouds
Changing Perceptions of the ISM




Densest regions detected first
Modeled as uniform “clouds”
Actually continuous spectrum of ρ, T, P.
Detection of motion showed dynamics
– Combined with early analytic turbulence models
– Success of turbulent picture limited then

Analytic tractability favored static equilibrium
models (or pseudo-equilibrium)
– Focus on heating/cooling, thermal phase transitions

New computational methods now bringing effects
of turbulence back into focus
Structure of Course
Lectures, Discussion, Technical
 Exercises
 Class Project
 Grading

– Exercises (30%)
– Participation (20%)
– Project (50%)
Project Schedule





Feb 24: Written proposal describing work to be
done (1-3 pp.). I’ll provide feedback on
practicality and interest.
Mar 10: Oral presentation of final project
proposals to class.
Apr 7: Proof-of-concept results in written report
(2-4 pp., including figures)
Apr 28: Oral presentation of projects to class in
conference format (10-15 minute talks)
May 5: Project reports due
Hydro Concepts

Solving equations of continuum
hydrodynamics (derived as velocity
moments of Boltzmann equation, closed by
equation of state for pressure)
D
  v  0,
Dt
Dv

 p  ,
Dt
De
     p v ,
Dt   
D 
where
  v
Dt t
where  2  4 G 
where p   kT / 
Following
Numerical
Recipes
Discretization
Consider a simple flux-conservative
advection equation:



 v    0, or in 1D
  vx
t
t
x
 This can be discretized on a grid of points in
time and space
x j  x0  jx

tn  t0  nt
Discretization of Derivatives

The simplest way to discretize the
derivatives is just FTCS:


t


x

 nj 1   nj
 O  t 
t
 nj 1   nj 1
2x
t
 O  x 2 
But, it doesn’t work!
x
Von Neumann stability analysis

The difference equation is
 nj 1   nj
t
  nj 1   nj 1 
 v 


2x



Suppose we assume

If |ξ(k)| > 1, then ξn grows with n exponentially!
 nj   neik ( jx) , where  
 e
n 1 ikj x

 e
n ikj x
t n ik ( j 1) x n ik ( j 1) x
 v
 e
 e


2x
Dividing by ξneikjΔx, and rearranging
t ik x
vt
 ik x
 1  v
e e
, so   1  i
sin k x


2x
x

|ξ(k)| > 1 for some k, so this scheme is unstable
Stability (cont.)

This instability can be fixed using a Lax scheme:
ρjn->0.5(ρj+1n+ ρj-1n) in the time derivative, so that
n
n





1 n
j 1
j 1
n 1
n
 j    j 1   j 1   vt 

2
2x



Now, if we do the same stability analysis, we find
1 ik x
t
 ik x
  e  e
v
eik x  e  ik x  ,


2
2x
vt
so now  ( k )  cos k x  i
sin k x
x
vt
2
2
and  ( k )  cos k x 
sin 2 k x,
x
v t
which is only  1 if
1
x
Courant condition

The requirement that
v t
x
 1is fundamental
to explicit finite difference schemes.

Signals moving with velocity v should not
traverse more than one cell Δx in time Δt.

Why is Lax scheme stable?
Numerical Viscosity

Suppose we take the Lax scheme
n
n

 j 1   j 1 
1 n
n 1
n
 j    j 1   j 1   vt 

2
2x


and rewrite it in the form of FTCS + remainder
n
n
n
n
n



 
 j 1   j 1
1  j 1  2  j   j 1 
 v 
 





t
2x
t

 2

n 1
j
n
j
This is just the finite difference representation of a
2
2


x





diffusion term
like a viscosity.
2


x  t2 
ZEUS
Program to solve hydro (and MHD)
equations (Stone & Norman 1992, ApJSupp)
 Details of numerical methods next time:

–
–
–
–
Second-order discretization
Eulerian moving grid
Artificial viscosity to resolve shocks
Conservative advection formulation
ZEUS organization

Operator splitting (Strang 1968):
– Separate different terms in hydro equations
– Source, advection, viscous terms each
computed in substep:
Take S   v . Then the momentum equation
dS
    vS   P  
dt
can be broken up into two steps
dS dS 
dS 




dt dt  source dt transport
ZEUS
flowchart

Timestep
determined
by Courant
criterion at
each cycle
ZEUS grid
Staggered grid
to allow easy
second-order
differencing of
velocities
 Grid naming
scheme…

Boundaries

“Ghost”
zones allow
specification
of boundary
values
–
–
–
–
Reflecting
Outflow
Periodic
Inflow
Version Control

Homegrown preprocessor EDITOR
– Clone of 70’s commercial HISTORN
– Similar to cpp with extra functions
– Modifies code two ways
• Define values for macros and set variables
• Include or delete lines

A few commands
–
–
–
–
*dk - deck, define a section of code
*cd - common deck, common block for later use
*ixx - include the following at line xx
*dxx[,yy]- delete from lines xx to yy, and substitute
following code
– *if def,VAR to *endif - only include code if VAR defined
File Structure
Baroque, to allow “automatic” installation
 From the top:

– zcomp, sets system variables for local system
– zeus34.s compilation script for ZEUS, EDITOR
– zeus34, source code with EDITOR commands
– zeus34.n, numbered version (next time)
– Setup block (next time) generates
• inzeus, runtime parameters
• zeus34.mac, sets compilation switches (macros)
• chgz34, makes changes to code
ZEUS installation
Copy ~mordecai/z3_template
 Run zcomp, wait for prompt. (First time takes
longer)
 View parameters, accept defaults, wait for
compile to finish
 Make an execution directory (mkdir exe)
 Copy xzeus34, inzeus into exe
 Run xzeus34. Progress can be tracked by
typing n

ZEUS output

To view output use IDL to read HDF files
Assignments

For next class read for discussion:
– Ferrière, 2002, Rev Mod Phys, 73, 1031-1066

Begin reading
– Stone & Norman, 1992, ApJ Supp, 80, 753-790
(I will cover more from this paper next time)

Complete Exercise 1
– Install ZEUS, begin reading manual, readme
files
– Begin learning IDL
– Review FORTRAN77 if not familiar