Download Reconciling Dwarf Galaxies with LCDM Cosmology Andrew Wetzel F RE

Reconciling Dwarf Galaxies
with LCDM Cosmology
Andrew Wetzel
Moore Prize Fellow
Carnegie Fellow in Theoretical Astrophysics
with the F RE collaboration
Feedback In Realistic Environments
dwarf galaxies present the most
significant challenges to the
Cold Dark Matter (CDM) model
Andrew Wetzel
Caltech - Carnegie
Context: Milky Way satellites and “crises” of ΛCDM
satellite dwarf galaxies: simulated vs observed
>105 identified subhalos
12 bright satellites (LV > 105 L )
V. Springel / Virgo Consortium
J. Bullock
Andrew Wetzel
Caltech - Carnegie
Kravtsov et al 2010
100
N(> Vcirc )
dark-matter subhalos in
dark-matter-only simulations
10
observed dwarf galaxies
around the Milky Way
1
10
20
30
40
Vcirc (km/s)
50
60 70
Vmax ⇡
r
GMtotal
r
“missing
satellites”
problem: too few observed satellites
ΛCDM
satellites
compared
with
dark-matter
subhalos
in
CDM
Local group dwarfs
Andrew Wetzel
Caltech - Carnegie
and Mateo, Olszewski & Walker 2
is analytically exact for sphericall
velocity dispersion profiles. Howe
systems the mass (circular velocity
by as much as 40 per cent (18 per cen
the long-axis, and similarly overesti
(22 per cent) if viewed from along t
2013). Shifts of order 20 per cent in
of the error bars on Draco and Ursa M
r
our overall conclusions. Other mass
GM2013;
r)& G
tot (<
Breddels
&
Helmi
Jardel
Vcirc (r)
=
are consistent with those
r plotted in
The points in Fig. 3 are sized by th
galaxy. Plotted in black are the lowmagenta points indicate the high-d
Minor, which may only be associate
host (indicated by the dotted magen
failures. If the data points for Draco
Boylan-Kolchin
2012
higher (e.g. if Vet
were
underestima
1/2al
Garrison-Kimmel
et vanish
al 2014
(cyan lines) would
but the nu
and black lines) would remain unch
Fig. 4 summarizes the results of
complete set of 48 hosts, where eac
“too3. big
fail”
problem:
dark-matter
in CDM
assumed density
profile are
shape. Bl
Figure
Rotationto
curves,
assuming
Einasto profiles
with α = 0.18, of subhalos
−1 within 300 kpc of the centre
fiducial
choice, an α = 0.18 Eina
all
resolved
haloes
with
V
>
30
km
s
peak
too
dense
compared
with
observed
satellite
galaxies
of Douglas (based on measured Vmax and Rmax values in the simulation).
implied distributions for NFW pro
Plotted as Wetzel
black points are the data for the MW satellites brighter than
Caltech
Carnegie
Einasto (cyan;
α =-0.28),
and an o
Andrew
dwarf galaxies: the most significant challenges
to the Cold Dark Matter (CDM) model
“missing satellites” problem
(probably) too few observed satellite galaxies compared with
dark-matter subhalos in CDM
—> Can a CDM-based model produce a satellite stellar mass
function as observed?
“too big to fail” problem
dark-matter subhalos in CDM are too dense compared with
observed satellite galaxies
—> Can a CDM-based model produce a satellite dynamical
mass (velocity dispersion) function as observed?
Andrew Wetzel
Caltech - Carnegie
The Local Group
Weisz et al 2014
Andrew Wetzel
Caltech - Carnegie
distance to MW or M31 [kpc]
Wetzel et al 2015
MW halo radius
~300 kpc
limited observational completeness of dwarf galaxies
Andrew Wetzel
Caltech - Carnegie
cosmological hydrodynamic simulations of
Milky Way-mass galaxy to z = 0
0.1
Latte
(better —>)
1
10
FIRE
100
1000
resolve dwarf
galaxies
GARROTXA
Agertz&Kravtsov
Sawala
Eris
CLUES
Mollitor
Massive MAGICC
GASOLINE
Eagle
Black-II Illustris
Aquarius (AREPO)
NIHAO
(better —>)
Andrew Wetzel
Caltech - Carnegie
The Latte Project: the Milky Way on FIRE
Simulating a Milky Way-mass galaxy with a realistic
population of satellite dwarf galaxies at parsec resolution
Wetzel et al 2016, ApJL submitted, arXiv:1602:05957
F RE
Feedback In Realistic Environments
F RE model for star formation
Feedback In Realistic Environments
http://www.astrophoto.com/M82.htm
Ultra-high resolution
mgas = 7000 Msun
hgas = 1 pc (hdm = 20 pc)
captures multi-phase inter-stellar medium
Cooling from atoms, molecules, and 9 metals down to 10 K
Star formation only in self-gravitating clouds: nH >~100 cm-3
Star formation efficiency: 100% per free-fall time
Andrew Wetzel
ESA
Caltech - Carnegie
F RE model for stellar feedback
Feedback In Realistic Environments
Heating:
Supernovae: core-collapse (II) and Ia
http://www.astrophoto.com/M82.htm
Stellar Winds: massive O-stars & AGB stars
30 Doradus
Photoionization (HII regions)
Explicit Momentum Flux:
Radiation Pressure
Ṗrad
L
⇠ (1 +
c
IR )
Supernovae
1
ṖSNe ⇠ ĖSNe vejecta
Stellar Winds
ṖW ⇠ Ṁ vwind
Andrew Wetzel
ESA
Caltech - Carnegie
.
n
n
t
f
r
h
s
a
Galaxy Stellar Mass [Mstar/fbMhalo]
a
s
e
n
t
e
e
f
l
no feedback
FIRE feedback
Hopkins et al 2014
log Halo Mass [Msun]Figure 4. Galaxylog
Halo Mass [Msun]
stellar mass-halo mass relation at z = 0. Top: M (M
∗
halo ).
Figure 6. M⇤ Mhalo relation at z = 0, as Fig. 4. Top: Simulations with
Bottom: M relative to the Universal baryon budget of the halo (fb Mhalo ).
different numerical parameters: we show the effects of varied resolution, ∗
Each simulation (points) from Table 1 is shown; large point denotes the most
artificial viscosity, and the algorithmic implementation of feedback. We
massive halo in each box. We compare the relation if all baryons became
also compare a completely different version of SPH (with a different set
stars (M∗ = fb Mhalo ; dotted) and the observationally inferred relationship
of hydrodynamic equations), which is known to differ significantly as
in determined
cerin Moster, Naab & White (2013, magenta) and Behroozi,
tain idealized hydrodynamic test problems. These have little effect Wechsler
on our & Conroy (2013, cyan) – dashed lines denote extrapolation beyond
predictions. Bottom: Effect of physical variation in stellar feedbackthepropobserved range; see footnote 9. The agreement with observations is
erties. We compare runs with no stellar feedback, with no supernovae
(but at Mhalo ! 1013 M , including dwarf though MW-mass galaxies.
excellent
relation between galaxy stellar mass and
host halo virial mass emerges naturally
Andrew Wetzel
⊙
Caltech - Carnegie
log SFR density [Msun/yr/pc2]
⌃˙ ⇤ ⇠ ⌃gas /⌧dyn
Galaxies on FIRE:
⌃˙ feedback
⇠ 0.02and⌃star formation
/⌧
⇤
gas
dyn
591
Hopkins et al 2014
d
e
erv
s
ob
log gas density [Msun/yr/pc2]
2
log
gas
density
[M
sun/yr/pc ]
Figure 8. KS-law, observed (Kennicutt 1998; Bigiel et al. 2008; Genzel
et al. 2010; Daddi et al. 2010, yellow shaded range) and simulated (points
as Fig. 4). We emphasize that this is a prediction: the instantaneous star
formation efficiency per dynamical time in dense gas is 100 per cent in the
simulations, but the emergent KS-law, as a consequence of feedback, has
an efficiency a factor ∼50 lower. As shown in Paper I and Hopkins et al.
(2013c), this is insensitive, with resolved feedback models, to the smallscale star formation law, and entirely determined by stellar feedback. ‘No
feedback’ models lie a factor ∼50 above the observations; ‘no radiation’
and ‘no SNe’ models (Fig. 6) lie a factor ∼10 above observations.
Kennicutt-Schmidt relation emerges naturally
Andrew Wetzel
Caltech - Carnegie
Latte: cosmological zoom-in simulation
dark matter
M200m=1.3x1012 Msun
6 Mpc
Latte: cosmological zoom-in simulation
stars
6 Mpc
dark matter only simulation
600 kpc
dark matter with effects of baryons
600 kpc
stars
600 kpc
Mstar = 9x1010 Msun
SFR = 3.4 Msun/yr
20
The Latte Project:
the Milky Way on FIRE
Population of
satellite dwarf
galaxies
F RE
Feedback In Realistic Environments
Andrew Wetzel
Caltech - Carnegie
stellar mass function of satellites
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
stellar mass function of satellites
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
stellar velocity dispersion function of satellites
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
stellar velocity dispersion function of satellites
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
velocity dispersion - mass relation
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
velocity dispersion - mass relation
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
mass - metallicity relation
Kirby et al 2014
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
mass - metallicity relation
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
diverse range of star-formation
histories of satellite dwarf galaxies
Weisz et al 2014
Andrew Wetzel
Caltech - Carnegie
diverse range of star-formation
histories of satellite dwarf galaxies
Wetzel et al 2016
Andrew Wetzel
Caltech - Carnegie
What causes the lack of (massive) satellite dwarf
galaxies around the Milky Way-mass host?
1. Stellar feedback forms dark-matter cores by
driving significant gas outflows/inflows that
transfer orbital energy to dark matter
2. Stellar disk of the Milky Way-mass host galaxy
destroys satellites (via tidal shocking, etc)
Andrew Wetzel
Caltech - Carnegie
inclusion of baryons destroys
dark-matter subhalos
dark matter in dark-matter-only dark matter in baryonic simulation
Andrew Wetzel
Caltech - Carnegie
subhalo number density profile
su
bh
alo
si
n
da
rk
-m
at
subhalos in baryonic
te
ro
nly
Mbound >1e7 Msun
Wetzel et al in prep
Andrew Wetzel
Caltech - Carnegie
dark-matter halo mass function
ha
lo
ha
los
in
s
in
da
rk
ba
ry
-m
on
ic
at
te
ro
nl
y
d >1000 kpc
Wetzel et al in prep
Andrew Wetzel
Caltech - Carnegie
dark-matter subhalo mass function
su
bh
a
lo
s
in
da
rk
-
m
su
bh
alo
at
s in
te
r
on
ly
ba
ryo
nic
d < 300 kpc
Wetzel et al in prep
Andrew Wetzel
Caltech - Carnegie
x
A Modest Proposal
“LCDM predicts…”
(dark energy + cold dark matter)
“LCDMB predicts…”
(dark energy + cold dark matter + baryons)
Andrew Wetzel
Caltech - Carnegie
The Latte Project:
the Milky Way on FIRE
F RE
Feedback In Realistic Environments
http://www.astrophoto.com/M82.htm
stellar mass
function
velocity
dispersion
function
mass - metallicity
starformation
histories
Andrew Wetzel
Caltech - Carnegie