Dwarf galaxies as gastrophysical laboratories

1
Dwarf galaxies as gastrophysical laboratories

• rachel somerville
• MPIA

ringberg march 2006
2
gastrophysics the big challenge

• in order to reconcile CDM with galaxy
  observations, there must be a strongly non-linear
  relationship between DM halo mass and baryonic
  mass
• achieved in semi-analytic models via scaling laws
  for star formation and feedback
• feedback from stars and AGN probably play a role
  -- relative importance still poorly understood!

DM halos
stars
3
merger history determines galaxy morphology
1.0E14
1.0E12
4
primordial power spectrum
BH formation, AGN feedback
galaxy observables
5
gas cools onto central galaxy... (satellites
dont get any new gas)
rcool
6
stars form and metals are produced (n.b. gas
recycling is included) cold gas (and metals)
heated and ejected from the galaxy either
retained in the halo or ejected entirely ejected
gas returns to the halo when larger scale
collapses
rvir
rturnaround
7
star formation

• dm/dt mcold/t0 tdyn(V0/Vc)a?

• t0 controls the fraction of gas turned into stars
• with ?0, equivalent to Kennicutt Law

(more efficient SF occurs in merger-triggered
bursts)
8
Martin-Heckman Law
SN feedback what do observations tell us?

• reheating rate few times SFR, no strong Vc
  dependence
• gas escapes from galaxies with Vc

Martin 1999
9
Efficiency of SN-driven outflows
detailed hydrodynamic simulations
mass ejection efficiency very low for Mg107
Mo blow-out only for very small masses, Mgejection of metals much more efficient
70-100 ejected for MgMac Low Ferrara 1999
10
MacLow Ferrara 1999
11
SN feedback and gas ejection
the standard semi-analytic treatment
reheated gas ejected from halo
-reheated mass times SFR typically assumed to be
a strong function of Vc -treatment of reheated
gas varies from model to model --
important! -here, reheated gas ejected from
halo when VcSilk see also Woo Dekel)
reheating rate/SFR
12
supernova feedback

• dmrh/dt b(Vc) dm/dt
• b(Vc) eSN (V0,SN/Vc)arh
• energetics arguments predict arh 2
• gas ejected from halo if VcVej
• ejected gas re-enters halo on mass doubling
  timescale

13
photoionization squelching

• after reionization, ionizing background prevents
  gas from collapsing in halos less than the
  filtering mass scale, 30-50 km/s

filtering mass
Gnedin 2000
14
Local Group Luminosity function
consensus that squelching solves the
substructure problem? what about the star
formation histories of dwarfs?
rss 2002
15
constraints on the (coldstar) baryon fraction of
halos
from size-mass relation of disk galaxies in
SDSS (Shen et al. 2003)
from Halo Occupation analysis (Kravtsov et al.
2004 N-body Zheng et al. 2005 semi-analytic)
16
no SN feedbackhot gas 0.3 solar
purple dots median baryonic mass
fraction yellow stars median stellar mass
fraction orange dotsbaryons, central
galaxies cyan dotsbaryons, satellite galaxies
Milky Way (Klypin et al. 2002)
K04 baryons K04 stars
17
model 0
Kennicutt SF law ?rh2, ?SN1, Vej100 km/s
purple dots median baryonic mass
fraction yellow stars median stellar mass
fraction orange dotsbaryons, central
galaxies cyan dotsbaryons, satellite galaxies
Milky Way (Klypin et al. 2002)
K04 baryons K04 stars
18
model 0
Kennicutt SF law ?rh2, ?SN1, Vej100 km/s
large box (L171 Mpc) smallest host halo 1011
Msun small box (L25 Mpc) smallest host halo 20
km/s
dominated by satellites
19
gas fraction vs. stellar mass
all galaxies
20
gas fraction vs. stellar mass
central galaxies
21
model 1
introducing...
tilted Kennicutt Law (?2.5) SN FB just like
model 0
central galaxies only
all galaxies
22
model 1
tilted Kennicutt Law (?2.5) SN FB just like
model 0
big box
small box
23
model 1r
tilted Kennicutt Law (?2.5) ?rh2, ?SN1,
Vej0 km/s
gas retained
gas ejected
24
model 2
tilted Kennicutt Law (?2.5) ?rh0, ?SN3,
Vej100 km/s
(based on Martin 99 recipe)
model 2 (small box)
model 1 (small box)
25
low mass galaxies have higher specific SFR than
massive galaxies
present/past averaged star formation rate
log stellar mass
Brinchmann et al. 2004
26
model 0
standard Kennicutt Law
central satellite
Brinchmann et al. 2004
problem too many low- mass galaxies with little
or no gas hence little or no SF
27
model 1
tilted Kennicutt Law
central satellite
problem remains now the galaxies have gas, but
their SF efficiencies are lower by design! SFE
function of mass z redshift (or
other variable)???
28
model 0
metallicity of cold gas
standard Kennicutt
Tremonti et al.
central satellite
big box
Lee et al.
observed dispersion for dwarfs 0.12 dex
small box
29
model 1
tilted Kennicutt
observed dispersion for dwarfs 0.12 dex
30
model 2
?rh0 (Martin99 FB)
31
model 1
model 0
closed box
Lee data

Garnett data
model 2
32
stellar metallicity distribution in MW M31
these scaling laws have consequences for big
galaxies too...
MW
M31
Rothberg, rss, Whitmore in prep
model 0
model 2
33
and, of course, they also change the redshift
dependence of all sorts of quantities
34
an aside reducing small scale power to get rid
of small DM halos has important consequences at
high redshift
M5x105 1/h Msun
T104 K
rss, Bullock Livio 2003
35
-RSI/WDM models have trouble producing enough
early star formation even at z2-6 -may have
trouble producing enough photons to
reionize the universe by z15 -fewer old stars
in galaxies at z1-3 -galaxy live in smaller
mass, less clustered halos
LCDM
RSI
36
where does this leave us?

• was star formation less efficient in dwarf
  galaxies in the past?
• threshold? ionizing background? metallicity?
• does mrh/dt /dm/dt scale with Vc?
• models seem to require this, but not seen in
  observations
• is there a critical velocity threshold for gas
  ejection, and where is it?
• Vej100 km/s seems to produce a break in the
  wrong place

37
what do we need?

• compilations of stellar mass, magnitude, circular
  velocity, gas content, metallicity, star
  formation indicators, size, etc, spanning the
  full range of galaxy masses from dwarfs to giants
• more information about dwarf properties at
  moderate and high redshift
• age metallicity distributions in giant and
  dwarf galaxies from resolved stellar population
  studies
• systematic, detailed comparisons with model
  predictions for all of these kinds of data

38
(No Transcript)
39
star formation SN feedback
??rh
reheated gas ejected if VcVej km/s

• major mergers (41) trigger bursts of star
  formation
• Bruzual Charlot 2003 multi-metallicity stellar
• population models w/Chabrier IMF (eq. to
  Kroupa/Kennicutt)

40
(No Transcript)
41
no SN feedback
purple dots median baryonic mass
fraction yellow stars median stellar mass
fraction orange dotsbaryons, central
galaxies cyan dotsbaryons, satellite galaxies
K04 baryons K04 stars