massimo ricotti (u of maryland) - astronomy · the first galaxies and the likely discovery of their...
TRANSCRIPT
The First Galaxies and the Likely Discovery of
their Fossils in the Local Group
Massimo Ricotti (U of Maryland)
Outline•
Introduction: Near field Cosmology
•
Part I : feedback‐regulated formation of the first galaxies and the quest for their fossils
•
Part II: Revisiting the idea of reionization
feedback on dwarf galaxy formation
The material presented here is collected in a recent review paper: Ricotti
2010 (arXiv0911.2792)
A lot of the work done by 2 PhD students:Mia Bovill
(on the job market next year)
Emil Polisensky
Collaborators:Nick Gnedin
(FNAL) and Mike Shull (Colorado).
The material presented here is collected in a recent review paper: Ricotti
2010 (arXiv0911.2792)
A lot of the work done by 2 PhD students:Mia Bovill
(on the job market next year)
Emil Polisensky
Collaborators:Nick Gnedin
(FNAL) and Mike Shull (Colorado).
The smallest halos and galaxies
1.
In CDM smallest dark matter halos have Jupiter mass
2.
First stars form in 105
Msun
mass halos
3.
Hierarchical formation: small DM halos and galaxies form first at z ≈
30‐50
4.
These small halos are still around to today: some merged, some got stripped, some
survived
The Missing Satellites
There should be thousands of dwarf satellites around the Milky Way.However, as of 2005 only about 30 satellites known in LG.Aquarius Simulation (Springel et al, 2008)
Klypin
et al 1999, More et al 1999Figure credit: Grebel
Less satellites if dark matter is warm
Assuming that all satellites are luminous:
Nsat
(WMD
simulation)≥Nsat
(observed)‐> we can set limits on warmness of DM
and thus the mass of WDM particle
This method is staring to become
competitive and complementary with
methods using Lyman‐alpha forest (eg,
Viel
et al 2006, 2008) due to the recent
discovery of many new MW satellites
Polisensky
& Ricotti
2010
ResultsThermal relic:mwdm
> 3.0 keV
at 68.4% CLmwdm
> 2.6 keV
at 95.4% CL
Sterile neutrino:ms
>18.9 keV
at 68.4% CLms
>15.7 keV
at 95.4% CL
Posisensky
& Ricotti
2010
2‐sigma
1‐sigma
Feedback regulated formation of pre‐ reionization
dwarfs
•
How small were the first galaxies?•
How many? … and how do they look like?
•
Where are the first galaxies today?
•
We do not have definite answers to these questions: not only the properties of the
galaxies but their number or even existence is determined by
feedback (radiative
and
chemical feedback)
Feedback on dwarf galaxy formation due to reionization
The smallest galaxies in the Universe can only form stars before reionization. But see part II of the talk!
The smallest galaxies in the Universe can only form stars before reionization. But see part II of the talk!
TIGM
=20,000 K
“pre‐reionization”
dwarfs
Are galaxies that DO NOT initiate gas condensation by Lyman‐alpha cooling. Roughly have:
Tvir
< 10,000 K or Mdm
< 108
Msun
or vcir
<20 km/s
Subject to feedback effects because in order to initiate star formation one of the following is necessary:
1.
Sufficient H2
formation in metal free gas (catalyst: H‐)2.
Sufficient metal pre‐enrichment
If none available these mini‐halos remain dark (forever?)‐‐> pre‐reionization
dwarfs DO NOT EXIST in this case
Important because number of mini‐halos ≈
1/M2
Simulations of the first galaxies
Feedback‐regulated
galaxy formation
•1 Mpc
box size•103
Msun
mass res.•3D radiative
transfer•Run ends at redshift
of
reionization
Ref: Ricotti, Gnedin
& Shull 2002
Pre‐reionization
dwarfs: numerous but faint!
Fractio
n of baryons in
stars
Pre‐reionizationdwarfs
How do pre‐reionization
dwarfs look like today assuming passive stellar evolution?
Ref: Ricotti
& Gnedin
2005
Blue ‐>
simulations
Black ‐> observationsof MW and Andromedasatellites as of 2005
The Sloan Digital Sky Survey Era
2.5 m telescope at Apache Point Observatory, New Mexico
Has covered ~20% of the sky down to ~ 22 magnitude with spectroscopy.
astronomers for scale
DR 7 sky coverage
About 25 new, ultra- faint Local Group dwarfs found in the last 5 years…and more are discovered every day!
Pre‐reionization
fossils and ultra‐faint dSphs: a model prediction!
Ref: Bovill
& Ricotti
09
Yet undiscovered Population?
Run D produces the observed distribution, but shows fewer high [Fe/H] dwarfs at ultra-faint luminosities.
See also Salvadori & Ferrara 2009
Budget of MW Satellites as of 3/2010
1.
Observations: correction for survey coverage of the sky and
completeness (see Touleroud
et al 2009, Walsh et al 2009, Koposov
et al 2009)
• Model independent correction: Nsat
>60 (Walsh et al 2009)•
Model dependent correction: 176< Nsat
<330 (Touleroud
et
al 2009)2.
Simulations: about 30
satellites with vmax
>20 km/s (non fossils) •
However, about 50‐60 satellites with vmax
<20 km/s where
more massive in the past (Kravtsov
et al 2006, Bovill
&
Ricotti, in prep)
Bovill
& Ricotti
2009
Hybrid Initial ConditionsThe final pre-reionization output is transformed in a 1 Mpc3 box of particles.
We duplicate this box, adding perturbations to account for density variations with l > 1 Mpc.
Each HR particle in the resulting N- body simulation represents a pre- reionization halo.
Unique IDs allow us to retrieve the stellar properties at z = 0.
Problem with our simulation: baryon physics?
Too many bright dwarfs in the outer
parts of the Milky Way
Bovill
& Ricotti
2010, in prep
Blue: true fossilsRed: polluted fossils+non
fossilsBlack: all satellites
Pre‐reionization
dwarfs: numerous but faint!Fractio
n of baryons in
stars
Pre‐reionizationdwarfs
Star formationefficiency is too high:1‐
SN feedback2‐
coef. in Schmidt law
The curious case of Leo T
•
Is a small dSph: σ*
=6.9 km/s. If it leaves in a halo with vcir
<20 km/s shouldn’t be able to accrete gas
•
Looks like all other pre‐reionization
fossils
•
Distance 430 kpc
from Galactic center
•
Is gas rich!
•
Forming stars at z=0!Leo T properties:� r
core
=100 pc �M
dm
/M
gas
=10‐30�N
H
=7x1020
cm‐2
Ref: Ryan‐Weber et al 2008
Leo T properties:� rcore
=100 pc �Mdm
/Mgas
=10‐30�NH
=7x1020
cm‐2
Ref: Ryan‐Weber et al 2008Can Leo T be a pre‐
reionization
fossil ?
Revisiting the idea of reionization feedback on dwarf galaxy formation
Ability of IGM gas to condense at the center of a DM halo increases with increasing:
1.
Γ=[vcir
/cs(IGM)
]2=Tvir
/TIGM
IGM temperature decreases at redshift
z<3
2.
Concentration of dark halo, c=rh
/rs
Concentration of first minihalos
increases with
decreasing redshift
if halo evolves in isolation: c≈
5
a/avir3.
Gas must be able to recombine (and cool)(note that tcool
<trec
)
Gas density profile in NFW potential
z=0, Γ=1
Late gas accretion onto mini‐halos
Ref: Ricotti
2009
Zvir
=10
Message to take home
•
The vanilla reionization
feedback idea fails at z<1 if the halo is highly concentrated
•
Minihalos
virialized
before reionization, that evolve to z=0 in the low‐density IGM, may
have a late phase of gas accretion and possibly star formation
13.5 10 7 4 0
AGE
SFR
13.5 10 7 4 0
AGE
SFR
Do these objects exist?
Dark Galaxies in Local Voids?
A future new test of CDM: gas rich minihalos
in nearby voids detectable in 21 cm?AFALFA survey may be able to find the most massive ones but we likely need deeperfuture surveys (with SKA) to probe the smallest minihalos.
Summary
New lower limits on WDM mass from MW satellites: mwdm
>2.6 keV
(thermal relic) and ms
>15.7 keV
(sterile neutrino) at 95.4% CL
(method independent of Ly‐alpha forest)
Neglecting the existence of an important population of pre‐
reionization
dwarfs (one of the most wide spread assumptions in
cosmology due to the complicated feedback physics required to
model these objects) is starting to appear unjustified by observations
!
Revisitation
of reionization
feedback: halos with vcir
< 20 km/s stop
accreting gas after reionization
but, if they evolve in isolation, may
have a recent (at z<1) phase of accretion from IGM and star
formation. This has two important implications:1)
Pre‐reionization
fossils may have bimodal SFH2)
Dark mini‐halos may accrete gas for the first time at z<1. Compact
HVCs
or undiscovered population?
Evidence 1: Evolution of the IGM temperature
(from observations of the Lyman‐alpha forest)
Ref: Ricotti, Gnedin
& Shull 2000
Evidence 2: Dark Halo Concentration
c(z) cvir1 zvir
1 z
5 1 zvir
1 z
Refs: Bullock et al. 2001, Wechsler et al. 2002, Ricotti
2003, Ricotti, Pontzen
& Viel
2007
Concentration increase as the halo evolves, after virialization, in expanding universe (low density IGM). From N‐body simulations:
Gas density profile in NFW potential
z=0, Γ=1
Mini‐halos with gas and no stars (comp. HVC?)
Ref: Ricotti
2009
Green: 4‐σ
pert.Red: 3‐σ
pert.Black: 2‐σ
pert.
Star forming mini‐halos at z=0 and z=2
Ref: Ricotti
2009
3‐σ
pert.
Compact HVCs
or unknown population?
If compact HVCs
are at distances of about 1 Mpc model may work for some of them
� rh
= 15 kpc� rcore
=100 pc
NH
=1018 ‐
1019
cm‐2
Refs: Blitz et al. 1999, Braun & Burton 1999, Robinshaw
et al. 2002, Sternberg et al. 2002, Putman et al. 2003, Maloney & Putman 2003,
Most are probably undetected because too faint and/or too small
More on properties of accreted gas
rcore 140pc vcir
17km/s
101 zvir
32
rh 18kpc M dm
10 8 M sun
13 at z = 0
M dm
M gas
5 30
vcir (rcore ) 0.66vcir 0.62vcirmax
Leo T: rcore
=100 pcMdm
/Mgas
= 10‐30NH
=7x1020
cm‐2
(Vcir
=11 km/s)
Leo T: rcore
=100 pcMdm
/Mgas
= 10‐30NH
=7x1020
cm‐2
(Vcir
=11 km/s)
To Fossilize or not to …1 – halo at z=0 has vmax < 20 km/s (red and blue lines)2 – when traced back to zinit , no component of the halo had vmax > 20 km/s (blue)
Introduction – Simulations – Distribution
20 km/s
0z
vmax