Star formation, feedback and the role of SNe II and SNe Ia in the
Download
Report
Transcript Star formation, feedback and the role of SNe II and SNe Ia in the
Three-dimensional hydrodynamical simulations
of ISM pollution by type Ia and II supernovae in
forming dwarf spheroidal galaxies
Andrea Marcolini
Fabrizio Brighenti
Annibale D’Ercole
(Bologna University)
(Bologna University)
(Bologna Observatory)
with the enormous contribution of:
Vienna 12/04/06
Simone Recchi (Wien University)
Francesca Matteucci (Trieste University)
Dwarf spheroidals galaxies of the local group were
originally thought to be very similar in their
metallicity and star formation histories to the galactic
globular clusters, but their star formation history is
now known to be much more complex.
Relative SFR
Draco
Relative SFR
Star Formation History in Dwarf Spheroidal Galaxies:
Time (Gyr)
Relative SFR
Relative SFR
Time (Gyr)
Sculptor
Time (Gyr)
Time (Gyr)
“The small dynamical mass of dwarf spheroidals means
that their binding energy is small compared to the energy
released by several supernovae, which leads the high
metallicity spread and relatively high mean metallicity
derived for these galaxies puzzling: how did the gas stay
bound enough to have an extended star formation and
gas enrichment?”
Babusiaux, Gilmore & Irwin 2005
Goals of the simulations
Find a galaxy model and a star formation history
reproducing the amount of stellar content in a
consistent way (without ejecting the whole ISM too
soon).
Reproduce the range of the observed metallicity.
Reproduce the mass-metallicity relation.
See if a supernovae feedback model (Dekel & Silk)
is able to explain the dwarf spheroidal (elliptical)
properties.
Assumptions for the Model:
Type
Iaparameter
Stellar
Type
Gas
IISupernovae:
Component
Component:
Supernovae:
:
Dark
Matter
Halo:
One
Star
formation
history
family
same
as Msol
before
following
&1992
Renzini
1983
time.
1 SNThe
IIKing
every
Model
Mgas
100
following
= 0.2*Mdark,
ofbut
formed
Peterson
in stars,
hydrostatic
&Greggio
Caldwell
uniformally
equilibrium
distributed
andwith
Mateo
T=Tvir
inintime
1998for 30 Myr
Modified
isothermal
obtained
1995 ranging
so that M/Lv
agrees
We
Weconstruct
choose
several
a one Halo
parameter
sequences
family
offollowing
instantaneous
modelsBurkert
of galaxies
bursts
differing
from
in number
10^6after each istantaneous burst. Stochastically distributed in space proportionally
with in
observations
(e.g.
Peterson
Caldwell
1992)
10^10
and
solar
intensity
masses.
such
The only
a way
parameter
that the
stellar
is the &
luminosity
mass
formed
(stellar
aftermass)
3 Gyr of
is the
to the stellar density.
always
galaxy.
the same.
Just a case: Draco
g cm^-3
g cm^-3
Density profile of the three component for both models
X axes (pc)
= Dark matter
= Gas component
= Stellar
component
M tot
80
Lv
Tvir 2103 K
X axes (pc)
M tot ( R )
80
Lv
M tot
220
Lv
Tvir 7 103 K
So high value of M/L are consistent with recent
observations…
•Kleyna et al 2002 based on stellar radial velocity dispersion found a
mass for Draco of 8 x 107 MO within 3 cores radii (=700 pc)
•Wilkinson 2004 based on 207 discrete stellar velocities inferred a
halo mass of 1 x 108 MO for Draco and 2 x 108 MO for Ursa Minor.
•Lokas 2002 find a M/LV=100 inside two Sersic radii and a total mass
typically of the order of 109 MO (M/LV=3500)
•Walker et al 2005 analyzing radial velocity of 178 stars in the Fornax
dwarf found that Fornax has a more massive dark matter halo than
found by previous studies (M/LV=10-40 in the sampled region) and
proposed models with M=108-109 MO
...and theorical works:
• Kazantzidis
2004 with high resolution N-body simulations showed
that models of dwarf galaxies with isotropic and tangetially
anisotropic velocity distributions for the stellar component fit the
data only if the surrounding dark matter halos has maximum
circular velocities in the range 20-35 km s-1
• Mayer et al made simulations of ram pressure stripping and tidal
interaction and find a best model for Draco with V=42 km s-1 .
• Mashchenko et al 2005 made N-body simulations of the formation
of dwarf spheroidal galaxies and found that the best fitting models
for Draco have a large range of DM halo virial masses (108-109 MO)
• Mashchenko et al 2006 made simulations of the interaction of Draco
with our Galaxy and ruled out the “tidal dwarf” hypothesis and find
that Draco is a cosmological halo with DM mass between 7 x 107 and
3 x 109 MO and the fraction of tidally stripped stars is < 3%.
M tot
80
Lv
M tot
220
Lv
M dark 2 107 M
M dark 6 107 M
M 0.56 10 6 M
M 0.56 10 6 M
M gas 3 106 M
M gas 11106 M
Ebind 111051 erg
Ebind 831051 erg
N SNII 5600
N SNII 5600
Different star formation history:
Nburst 10
N SNII 560
t sleep 300Myr
N burst 25
N SNII 224
t sleep 120Myr
N burst 50
N SNII 112
t sleep 60Myr
}
t SFH 3Gyr
(Myr)
= 1 burst
= 10 burst
= 25 burst
= 50 burst
results for model with a total M/L=80
Time evolution of the gas mass inside the dark matter halo
= 50 bursts
= 25 bursts
= 10 bursts
= instantaneous
starburst
M (t )
M ( 0)
Time (Myr)
Independent of the star formation history the galaxy loses its gas in a short period
compared to the observed one (3 Gyr) and a few percent of stars are formed.
M tot
220
Lv
N burst 50
N SNII 112
t sleep 60Myr
Ebind 831051 erg
Time evolution of the gas mass inside the dark matter halo
Inside dark
matter halo
(1.2 kpc)
Inside stellar
region (600 pc)
More that 60% of the gas is still inside the dark
matter halo, and is bound to the galaxy for this model
SN II ejecta evolution
M
SN I ejecta evolution
M
time (Myr)
time (Myr)
= period with SN II
= total mass ejecta
= mass ejecta inside RDM ,cut
= mass ejecta inside R, cut
{
SNII
SNIa
M
Mej 101M
.4M
ej
M ej , Fe 0.07 M
M ej , Fe 0.7 M
M ej ,O 1.0 M
Inside dark matter halo
Inside stellar region
69 % of the total
amount of Fe
ejected after 3 Gyr
come from SNIIs
60 % inside the
galaxy region
15 % inside the star
forming region
31 % of the total
amount of Fe ejected
after 3 Gyr come
from SNIs
70 % inside the
galaxy region
17 % inside the star
forming region
Time evolution of the metallicity of the stars formed during the
various bursts of star formation
Time evolution of the metallicity of the stars formed during the
various bursts of star formation
[ Fe / H ] 1.65
Fe / H 1.5
Only SN II
SN II plus
SN I
[ Fe / H ] 2.0 0.21
[ Fe / H ] 1.8 0.20
[ Fe / H ] 1.7 0.10
e.g. Mateo (1998), Shetrone, Côtè & Sargent (2001)
Aparicio, Carrara & Martinez-Delgado (2001)
Bellazzini, Ferraro & Origlia (2002)
model with 50 bursts
model with 25 bursts
{
{
SNIa
SNII
M ej 1.4 M
M ej , Fe 0.74 M
M ej ,O 0.15M
M ej 10 M
M ej , Fe 0.07 M
M ej ,O 1.0 M
Shetrone et al (2001):
[ Fe / H ] 2.0
Our models
[ Fe / H ] 1.7 / 1.6
If the explosion of a
SNIa occurres during
the re-collapse phase
its ejecta results to be
much more localized.
ISM density
SNII ejecta density
SNIa ejecta density
A high density region is
formed, where the
conditions for a new
star formation episode
is recovered, which
has the dimension of
the stellar region
At the beginning of the
simulation there is a high
spread in metallicity, while
at later times the SNII
ejecta becomes more
uniform.
.... Ram pressure stripping and tidal interaction?
Marcolini, Brighenti & D’Ercole made simulations of disk-like
dwarf in poor groups and found that ram pressure can remove
most of the gas in dwarf galaxies with halo circular velocities
lower than 30 km s-1.
Mayer et at 2006 simulated the interaction of dwarf spheroidal
galaxies with the Milky Way halo and found that a galaxy similar
to Draco can be stripped completely of its gas in a time scale 2-3
Gyr if the gas is maintained at 104 K by some mechanism, a
timescale comparable to our assumptions for the star formation.
We are running now 2D simulation of the interaction of our model
with the Milky Way Halo and find similar results.
Conclusions and
Future work
•
If a “low” M/L (~80) is assumed the gas is blown away after few star
formation episodes, it doesn’t matter how weak they are.
•
If we assume an extended dark matter halo and weak star formation a
plausible scenario is possible, where the energy of the supernovae is
radiated away and the galaxy is able to retain its gas for a period
compatible with the observed star formation history.
•
With a model M/L =280 (within 1.2 Kpc) we are able to simulate in the
typical enrichment and the metallicity spread observed for this galaxy
(Draco).
•
We will run simulations in which the interaction of the galaxy with the Milky
Way halo is taken consistently into account and with a self consistent star
formation history.
•
We’d like to extend this model at higher masses (108-1010 Msol) and try to
simulate the mass-metallicity relation. We’d like to run models for galaxies
like Sculptor, Fornax, NGC147, NGC5206.