How Do Galaxies Get Their Gas

1
How Do Galaxies Get Their Gas?

• astro-ph/0407095
• Duan Kere
• University of Massachusetts
• Collaborators
• Neal Katz, Umass
• David Weinberg, Ohio-State
• Romeel Davé, University of Arizona

2
Standard Model

• White and Rees 1978
• Gas infalling in dark matter halo, shock heats
  to the virial temperature at the virial radius,
  and forms quasi-hydrostatic equilibrium halo.
  
• Shocked, hot, gas slowly cools and travels
  inwards, forming the central cooled component
  the galaxy.
  
• Basis for Semi-Analytic models (White and Frenk
  1991, Cole et al.1994, Somerville and Primack
  2000 etc.).

3
Less Standard Results

• Binney (1977) showed that only small part of the
  gas in his simulation reached TTvir
  
• Katz and Gunn (1991) most of the gas in
  simulation of a single galaxy formation never
  reached high temperatures ( radiated mostly in
  Lya).
• Kay et al (2000) only small fraction of the
  accreted gas in their simulated galaxies was
  heated to Tvir.
  
• Fardal et al. (2001) most of the cooling
  radiation from the halos is radiated in Lya with
  T virial temperature).

4
Simulation Properties

• We use PtreeSPH (Davé et al. 1997), with gas
  physics and star formation based on Katz et al.
  1996.
  
• Main simulation 22h-1Mpc on a side, 1283 SPH
  particles and 1283DM particles.
  
• Group (galaxy) resolution 6.8x109 M?.
• Higher resolution simulations are used to check
  numerical effects.

5
SFR, Accretion, Mergers

• Galaxies grow via mergers and gas accretion.
• Smooth accretion, not mergers, dominates the mass
  growth of galaxies! (Murali et al.
  2002, Kere et al. 2004)
  
• SFR tightly follows smooth accretion rates.

6
Checking the standard model...

• We track the accreted particles backwards in time
  and register maximum temperature gas particle had
  before it was accreted by a galaxy.
  
• In standard model this Tmax should be close to
  virial temperature.

7
Results

• Distribution of maximum temperature reached by
  the gas is clearly bimodal (Katz et al., 2002).
  
• Cold mode 104-105K
• Hot mode106-107K
• Local minimum in the Tmax distribution is at
  2.5x105K.
  
• Cold accretion mode is important at high
  redshifts and hot mode at low redshift.

8
Tmax/Tvir ratio also show clear bimodality
9
Global history of the cold and hot accretion

• Cold mode dominates at z3, while hot mode
  dominates at z
• Both modes drop significantly from high to low
  redshift.
  
• When integrated over the cosmic time both modes
  contribute similarly to the total accretion.

10
Dependence of the accretion rate on the
environment

• High z both modes important in all
  environments
  
• Low z hot mode dominates in high density
  regions, cold in low density regions.
  
• Accretion (SFR) drops in high density regions.

11
Cold/hot fraction dependence on the parent halo
mass

• Cold mode dominates in low mass halos
• Hot mode dominates in massive halos.
• Transition between modes 2 x 1011 M? .
• Similar to the model and 1D sims of Birnboim and
  Dekel (2003) !
  
• Virial shock does not develop in low mass halos
  due to fast cooling.

12
Cold/hot fraction dependence on galactic mass

• Transition at Mgal 2-3 1010 M?.
• Larger dispersion
• Small satellites hot mode dominated in massive
  halos.

13
Comparison with observations

• We closely mimic observational selection.
• Sharp drop in SFR happens at S1Mpc-2 or out of
  virial radius.
  
• SFR high in sims. but shape of the relation is in
  qualitative agreement.
  
• No need for gas stripping from the disc,
  truncation (strangulation) of gas supply
  qualitatively re-produce observations.

Goméz et al. 2003
14
Robustness

• Test with higher resolution simulations bimodal
  distribution of Tmax is not a consequence of the
  limited resolution (details in Kere et al.
  2004)
• High resolution 1D simulations of Birnboim and
  Dekel shows qualitatively the same result!
  
• A. Kravtsov Eulerian simulation (much better
  shock capturing) show qualitatively same result!
  
• Initial tests of GADGET simulations (Hernquist
  and Springel 2002), show similar bimodal gas
  accretion!
  
• Existence of cold and hot accretion is real
  physical result not just numerical artifact!

15
Accretion geometry
ri
rj
ri
rj
cos ri rj
-1
-1
cos ri rj
1
1
16
(No Transcript)
17
z5.52
z3.24
19,000p
90,000p

• Green accreted gas.
• Left COLD MODE (z5.52, M2.6x1011M?)
• All halo gas in the filaments.
• Cold gas, no virial shock
• Directional accretion
• Right HOT MODE halo (z3.24, 1.3x1012M?)
• Filled with gas, quasi-spherical
• TTvir
• Cold filaments penetrating the halo

r
r
2Rvir
-2Rvir
0
T
T
T
0
-0.5Rvir
0.5Rvir
18
Hot Mode

• Temperature much different from the virial
  temperature at the virial radius, grows inwards.
  
• More standard at lower redshift (more massive
  halos).
  
• T_vir reached only in the inner 0.5 Rvir.
• Large dispersion around Tmax/Tvir1
• Filaments penetrate deeply inside hot mode
  dominated halos (see also Nagai Kravtsov 2003)
  
• Temperature profile
• Numerical effects

19
Cold Mode

• Virial shock never develops in halos with M 1011 M? (Birnboim and Dekel) due to the short
  cooling time (unstable shock), instead cold gas
  gets to galaxy in free fall time!
• Cold mode is highly filamentary.
• All the halo gas is in galaxies or filaments
• Slowing in a series of small shocks, vrad drops
  inwards.
  
• Small hot halo around the galaxy some
  shocking at the disk.

20
Why bimodality?

• In small halos virial shock fails to develop, due
  to short cooling time (Birnboim Dekel 2003).
  There is no virial shock, since there is no hot
  halo gas accumulated in the halo !
• No hydrostatic equilibrium, collapse on the
  dynamical time scale (hints also in White Frenk
  1991).
  
• Big halos developed virial shock,
  quasi-hydrostatic equilibrium, gas spends some
  time in the halo before it cools.
  
• Much more complicated, temperature gradients,
  filaments in the halos!

21
Some Consequences

• Filamentary cold mode accretion can have
  important consequences on the angular momentum
  acquisition.
  
• Much of the dissipated energy from the galaxy
  formation at high z is emitted in Ly? (only some
  part in X-rays).
  
• Observational evidence Matsuda et al. 2004.
• Feedback fundamental difference in two modes.
• In cold mode halos gas is only in the filaments.
  Winds from galactic supernovae can easily expel
  the gas out of the halos, while it is hard to
  stop the gas falling in.
• Hot mode halos are full of hot gas that moves
  slowly, hard for winds to expel the gas but
  much easier to heat up the gas and slow or stop
  cooling (for example with AGNs)

22

• Stopping or slowing the hot accretion by some
  sort of energy input will allow large galaxies to
  grow mainly via mergers
  
• much better agreement with observed morphological
  types, colors and bright end of LF.
  
• expelling gas out of cold mode galaxies will make
  better agreement on the low end of the LF.
  
• Galactic properties (metallicity, SFR, galaxy
  type etc.) change at 1-3x1010M? (Kauffmann et
  al. 2003, Tremonti et al. 2004) suggesting that
  existence of the two accretion modes and their
  different properties may be responsible.

23
Summary

• 1. Smooth accretion is a dominant process of gas
  supply to the galaxies (not merging).
  
• 2. Roughly half of the gas accreted onto galaxies
  does not shock at virial radius.
  
• 3. Filamentary cold accretion dominates in halos
  with Mhalo
  2x1011M? hot mode dominates.
  
• Similar results from other simulations
  (1D-Birnboim and Dekel, Eulerian-Kravtsov et al.,
  GADGET-Springel and Hernquist).
  
• 4. Existence of separate cold and hot accretion
  modes can lead to interesting theoretical and
  observational consequences.
  
• For SFR-density dependance and other interesting
  details check Kere et al. 2004, astro-ph/0407095