Dwarf Spheroidal Galaxies : From observations to models and vice - - PowerPoint PPT Presentation

Dwarf Spheroidal Galaxies : From observations to models and vice versa

Yves Revaz

The good reasons to study dSphs :

Test for the LCDM paradigm : small structures are predicted in abundance Bower et al. 2003 Galaxy luminosity function

The good reasons to study dSphs :

The missing satellite problem LCDM simulations predict the existence of a high number of dwarf galaxies orbiting around Andromeda and the MilkyWay... ...but only a modest population is observed.

Simon & Geha 2007 Via Lactea (Diemand et al. 2007)

The good reasons to study dSphs :

In the hierarchical LCDM paradigm, dSphs are the building blocks of larger objects

Dwarf galaxies Spiral galaxies Elliptical galaxies Galaxy clusters

time

The good reasons to study dSphs :

Much more complex than usually thought Tinguely 1970

Agenda

What is a dSph galaxy ? How do we model these objects ?

➢ The driving parameters ➢ Intrinsic or extrinsic evolution ?

Can we predict the stellar mass for a given halo mass Some conclusions

What is a dSph galaxy ?

Faint stellar system (Mv < -15) Low velocity dispersion (~10 km/s) Evidence for dark matter No clear rotation (dominated by random motions) No ongoing star formation No gas (or at least not detected) [Fe/H] follows luminosity

Mateo 2008

dSphs are interior to transition systems, themselves interior to dIrrs

clustered around spiral galaxies

core tidal core tidal

diffuse objects

Fornax Carina Piatek, 2007; 2003

they form a sequence

Adapted from Tolstoy, Hill & Tosi 2009

Vt=198 ±50 km/s Vr=79 ±6 km/s Peri : 68 kpc (33, 83) Apo :122 kpc (97,133) Period : 2.2 Gyr

Sculptor

Piatek et al. 2006, 2003

Vt=85 ±39 km/s Vr=22 ±24 km/s Peri : 20 kpc (3, 63) Apo :102 kpc (102,113) Period : 1.4 Gyr

Carina

today today

...but are also very different

Sculptor Carina

Time Time Star Formation Rate

Rizzi et al. 2004

Out In

also Tolstoy et al. 2004; Battaglia et al. 2006

disparate, structures, diff. stellar pop as fct of r

What is at the origin of the diversity ?

• Is the environment the driving parameter ?

➢ Chance encounters produce a variety of properties

• Can we think otherwise ?

➢ How much of the variety is intrinsic ? ➢ When and how is interaction required ? ➢ Sequence = single framework of formation ?

Current limitations

➢ Too simplistic chemical evolution ➢ Arbitrary fixed SFH ➢ Small number of simulations ➢ Evolution stopped at high z ➢ Confront results only with global relations

Global relations are not all...

Objects with very similar :

• Lv
• [Fe/H]
• [M/Lv]

...

... may have completely different stellar and chemical properties !

SFR, M* AGE [Fe/H] [Mg/Fe]

Observations : Where do we stand ?

• Accurate abundances measurement from individual stars for:

➢ Fornax

(Letarte et al. 2010)

➢ Sculptor

(Hill et al. in prep.)

➢ Sextans

(Shetrone et al., 2001, Aoki et al. ,2009, Jablonka et al., in prep.)

➢ Carina

(Koch et al. , 2008, Lesmale et al. , 2011, Venn et al., in prep.)

DART Team DART Team

Modelisation

Code & physical processes The initial conditions Tests and Robustness

GEAR : a self-consistent Tree/SPH code (Revaz & Jablonka 2012)

Skeleton : Gadget-2 (Springel 2005) Gravity :

• > treecode (Barnes & Hut 86)

Hydrodynamics :

• > SPH : Smooth Particles Hydrodynamics (Lucy 77, Gingold & Monaghan 77)

NlogN instead of N2

Hydrodynamics values are obtained by convolution of neighbors particles with a kernel function

the resolution follows the mass

GEAR : a self-consistent Tree/SPH code

The baryon physics

(Katz 92) Cooling function (metal dependent) :

• above 104K (Sutherland & Dopita 93)
• below 104K, H2, HD, OI, CII, SiII, FeII

(Maio et al. 07)

Star formation : classical recipes : Schmidt law (Schmidt 59) + star formation density (0.1 a/cm3) + starformation temperature (3x104K)

Chemical evolution SSP scheme :

(Poirier 03, PhD thesis)

• SNIa, SNII nucleosynthesis + feedback from SN explosions
• elements followed : Fe, Mg

GEAR : a self-consistent Tree/SPH code

Time Formation time

• f a stellar

particle (SSP) mass fraction yields energy

IMF : Krupa 01 Kodama & Arimoto 97

Injection into the system : nearest neighbors

SNII : yields of massive stars (Iwamoto et al. 99) Energy : eSN 1051 ergs/SN (thermal)

SNIa : model from Kobayashi et al. 2000 yields from Tsujimoto et al. 95, updated models Energy : eSN 1051 ergs/SN (thermal)

stellar particles :

➢ Fe, Mg, Z, age

• > Lv (Vasdekis et al. 96)

gas particles :

➢ Fe, Mg, Z, density, temperature

all particles :

➢ positions, velocities, mass

Outputs of models and observables

Initial conditions

Initial conditions

• 2 Mpc/h3 box, dark matter only (WMAP V)
• 134'217'728 particles

=> resolution of 150 pc/h, 4.5 104 Msol/h more than 150 dSph haloes with masses between 108 > M > 109 Msol

Z=6.53 3.72 2.46 1.72 1.46 0.25 0.0

Z=6.53 3.72 2.46 1.72 1.46 0.25 0.0

50% of the haloes have their density profiles already in place at z=6 (in physical coordinate) 98% have an NFW profile, wich central densities varying by a fractor ~3 only (for masses between 108 and 109 Msol)

• Models of dSphs in a static Euclidean space, where the expansion of the universe is neglected,

are justified. The physics of baryons that depends on the density in physical coordinates is correct.

• The densities of haloes with mass between 108 and 109 Msol exhibit a small dispersion, a factor

3 to 4, which could help understanding the variety in the observed properties of the dSphs Core profile supported by the observations (Blais-Ouellette et al. 2001; de Blok & Bosma 2002; Swaters et al. 2003; Gentile et al. 2004, 2005; Spekkens et al. 2005; de Blok 2005; de Blok et al. 2008; Spano et al. 2008)

Initial conditions : isolated models

Energy conservation, Robustness & Convergence tests

Conservation better than 5% over more than 250 dynamical times !

9.5x108Msol

chemical properties :

• metallicity distribution
• abundances

density profile

Results

Parameters (see also Revaz et al. 2009)

More than 400 simulations, Exploring the effect

• f the parameters

Efficiency of supernova energy :

➢ If eSN=100% (1051 ergs per SN) → no Fe/H enhancement (need to be below 10%) ➢ no strong winds : the gas is kept around the system ➢ 107 Msol of gas linked to the dSph (see also Marcolini et al. 06, Valcke et al. 08)

eSN= 100% eSN= 1%

Mass = 4x108Msol

Mass or density ?

mass density

density x 5 : Δ[Fe/H]= 1.0 dex mass x 9 : Δ[Fe/H]= 0.5 dex

• Cooling stronger for larger densities
• Mass increases luminosity but has
• negligible impact on the chemical
• evolution

c*=0.05 eSN=0.03

Fornax Sculptor Sextans Carina

7x108Msol 5x108Msol 3x108Msol 1x108Msol

Yes

• The dominant driving parameters are the mass and density

(compatible with the cosmology)

• More massive and dense systems, forms stars continuously

➢ → high [Fe/H]

→ high Lv

• Less massive and less dense sytems forms stars episodically

➢ → low [Fe/H]

→ low Lv But we need outer physical processes

➢

to truncate the star formation

➢

to get rid of the remaining gas

Do we observe a sequence ?

Metallicity gradients

Sculptor (Tolstoy et al. 2004) Sextans (Battaglia et al. 2011) Fornax (Battaglia et al 2006)

Metallicity gradients

metalicity gradients ?

Metallicity gradients : the effect of gas motion

High resolution model : M=3.5 108Msol, 4x106 of particles

➢ hot gas heated by SNs accumulates at the center ➢ due to strong Archimedes forces this gas is driven outside ➢ high metallic gas is ejected into the IGM (1.5 105 Msol)

High resolution model : M=9.5 108Msol, 4x106 of particles

Metallicity gradients : the effect of gas motion

How galaxies populates their dark matter halo ?

Abundance matching

• Millennium Simulation (MS; Springel et al. 2005)
• High resolution MS (Boylan-Kolchin et al. 2009)
• SDSS/DR7

Assume

• main subhaloes and satellite subhaloes have

galaxies at their centres,

• the stellar masses of these galaxies are directly

related to the maximum dark matter mass ever attained by the subhalo during its evolution. In practice, this mass is usually the mass at z= 0.

• one-to-one and monotonic relationship between

Mhalo and M*

n(>Mhalo) = n(>M*)

Guo et al. 2010

SDSS+Millenium Guo et al 2010 Sawala et al 2011 Extrapolation Sawala et al 2011

Abundance matching :

Sawala et al 2011 SDSS+Millenium Guo et al 2010

Who is right ? Who is wrong ?

Abundance matching

SDSS+Millenium Guo et al 2010

Faber-Jackson / Tully-Fisher relation

SDSS+Millenium Guo et al 2010

Faber-Jackson / Tully-Fisher relation

SDSS+Millenium Guo et al 2010

Faber-Jackson / Tully-Fisher relation

Can we find 10 km/s stellar systems in 1010Msol halos ?

Large sample of halo populated by stars with random properties (N-body systems)

• Halo : generalized NFW (HonhSheng Zhao 1996, Treu et al. 2002) CUSP or CORE

Mh : [108,1011] Msol rs,h : [0.5-4] kpc γh : [0,1]

• Stars : generalized NFW

Mh : [105,109] Msol rs,s : [0.5-4] kpc γs : [0,1] Velocities computed from the Jeans Equations

Carina-like Fornax-like

Baryonic fractions at odd with plausible star formation histories

bf : [3x10-4, 3x10-3] bf : [3x10-5, 3x10-4]

Maccio et al. 2006

Conclusions

• GEAR : a self-consistent Tree/SPH code for dSphs simulations

➢

robust + good convergence

• In LCDM, the density profiles of small haloes does not evolve much betweeen z=6 and z=0

→ Possible to study dSph in an isolated context

• To fit the metallicity of dSphs, the feedback efficiency cannot be large :

➢

no strong winds but a large amount of gas left

• Global relations are reproduced
• But also the variety in stellar populations and chemical abundances for :

Fornax, Sculptor, Sextans & Carina

Conclusions

• Evolution driven by :
• Intrinsic processes : mass+density
• Extrinsic processes : get rid of the remaining gas + truncate the SFR
• The low stellar content predicted by LCDM for 1010Msol seems to be incompatible with our models
• Tidal stripping ?
• Ram pressure stripping ?

Credit : Olivier Tiret

The futur of GEAR

Cosmological simulations dIrr and spiral galaxies

The End