Does Gas Cool From the Hot Phase?(onto galaxies)

The MPA/ESO/MPE/USM 2007 Joint Astronomy Conference

Gas Accretion and Star Formation in Galaxies

Joel Bregman (Univ. of Michigan)

Hot Gas in/around Ellipticals: Lessons Learned

Hot Gas Environment of the Milky Way and Local Group: Limits on Accretion

Missing Baryons in Galaxies and Galaxy Groups

Let’s Look at Galaxies with a lot of Hot Gas that is Cooling

• Gas masses of 1E8-1E10 Msun• T = 3-8 E6 K; Lx ~ 1E41 erg/s• Cooling rate of 0.1-1 Msun/yr

This is known as

Cooling Flows

X-ray contours

How to Speak “Cooling Flow”

• Two types of hot gas situations– Hot gas in clusters of galaxies– Hot gas in early-type galaxies

• Clusters of Galaxies – Most of the baryons are gaseous (not galaxies)– 2-10x107 K– Cooling: Free-Free (X-Rays)– Last Millennia: cooling rates 100-1000 Msun/yr– This Millennia: cooling rates of 1-30 Msun/yr

The last Cooling Flow Meeting

Cooling Flow Ellipticals

• Most Ellipticals are not X-ray bright– Bright ones in groups/clusters (1041 erg/s)

• “Lots” of hot gas (~109.5 Msun)• Gas is bound• X-rays mainly due to line emission

– X-ray faint galaxies (1040 erg/s)• LMXBs + a little hot gas (~108 Msun)• Galactic Winds

• Bright Ellipticals are well-studied

Metallicity of the Hot Gas

X-Ray Observations (XMM-Newton)

Fe is about Solar ([Fe] = 0)

[O/Fe] = -0.3

Metallicity like stars

Not like Cluster/group [Fe] = -0.5

Gas from Stellar Mass Loss

Not dominated by accretion from cluster/group

XMM RGS spectrum of NGC 4636 (Xu et al. 2003)

Does this Gas Cool?• Observe OVIII (hot, ambient;

5E6 K)• OVI often detected in X-ray

bright galaxies– 3E5 K gas; evidence for cooling

from hotter gas

Far Ultraviolet Spectroscopic Explorer (RIP)

OVI is detected in about 40% of galaxies

Some cooling gas at 3x105 K

Cooling rates of 0.1-0.5 Msun/yr

In central region; not distributed.

Bregman et al. (2005)

Where Does The Cooled Gas Go?

• Es in the RSA Catalog (Hogg et al. 1990; Bregman et al. 1992)

• 104 K gas detected – Same metallicity as stars and hot gas– Not much mass (< 1E5 Msun)

• HI and H2 rarely detected– M(HI) < 3E7 Msun for many ellipticals; 5% detection rate– M(H2 ) < 2E8 Msun; 0% detected– Limits the mass of HI HVC

• Low levels of star formation, if present– Probably less than 0.1 Msun/yr

Lessons Learned From Ellipticals

• Ellipticals appear to have much more hot gas than Spirals (total ISM mass similar)

• Cooling flows in Es don’t lead to detectable masses of cooled HI (< 3E7 Msun)

• Not much evidence for accretion from surrounding group/cluster medium

• Radially distributed cooling should not occur– Local Thermal Instabilities suppressed in near-

hydrostatic situations (Balbus 1995)– Even most non-linear disturbances are suppressed

(Reale et al. 1991)

Hot Gas Around Spirals and in Galaxy Groups

• Stellar evolution models for the Galaxy– Need to resolve the G-dwarf problem (Z > 0.2)

– Accrete 1-2 Msun/yr of gas with [Fe/H] < -3

• Lx = 4E40 (Mdot/1Msun/yr) (T/3E6 K) erg/s

• Milky Way has Lx ~ 1039.3 erg/sec

• Other similar spirals have Lx ~ 1039.5 erg/s

• No obvious support for accretion with rates exceeding 0.1 Msun/yr

NGC 891 in X-raysChandra ACIS-S; 108 ksec of cleaned data

Point sources removed; smoothed

Hot gas easily seen to a height of 4.5 kpc from disk (1.6’) Fainter emission goes to a height of 9 kpc (3’)

Extent along disk similar to Halpha, FIR

Oosterloo et al. 2007; NGC 891 HI

Chandra 0.3 keV gas; 8’ box

Thermal emission; 4E39 erg/s

Mdot = 0.1 Msun/yr

Give Up The Idea Of Accreting Pristine Gas Onto MW

• Every Galaxy Group with good X-ray data– 0.0 > [Fe/H] > -0.7

• Sightlines out of the MW show OVII and OVIII in absorption (metals are present)

• Need to adopt a different approach to solving the G-dwarf problem– pollute the gas with metals before accretion

(Binney and Merrifield; Galactic Astronomy)

A Census of 106 -107 K Gas in the Milky Way and Local Group

Z = 0.1 Solar

N = 1019 cm-2

OVII, OVIII have X-ray resonance lines; best sensitivity

Detect Hot Gas by X-ray Absorption Lines

Galactic Halo Model: distribution largely spherical around the MWcolumn densities similar in all directionsmight see evidence for the shape of the Galaxy

Local Group ModelLocal Group is elongated along the MW-M31 axis

columns greater along this axis and especially toward M31 (the long way through the LG)

Concern: M31 may have its own extended halo

Discriminating Between a Galactic Halo and Local Group Model

Group simulation like the Local Group (Moore) shows elongation of matter .

Column enhancement along major axis ~2-3x perpendicular to axis

26 Target AGNs; mean EW = 22 mÅ; 17 with rms < 10 mÅ

These are the 4 best.

Bregman and Lloyd-Davies (2007; arXiv:0707.1699)

Toward Bulge

An AGN toward M31 (long axis of Local Group)

One of the smallest columns, not one of the largest

Local Group Model Prediction

The OVII data don’t fit a Local Group Model

Prediction of Local Group Model

A Galactic (Halo) Model Works Better

Central Line Velocities close to MW

Correlation with Galactic ¾ keV X-Ray Background

Supports a Galatic Halo Origin (10-100 kpc)

Gas Mass of 108 – 1010 Msun

Compare to NGC 5746 (Rasmussen & Ponman 2004): 109 Msun for similar metallicity

Other Evidence for Hot Gas In Local Group

• Nearby LG dwarfs have no gas but distant ones have gas (Blitz & Robishaw 2000; Grcevich et al. 2007)

– Ram pressure stripping – n = 2.5E-5 cm-3 at d = 200 kpc (less than

column inferred from OVII line by 4x)

– 1E10 Msun of gas

– Cooling time longer than Hubble time

X-Ray Shadowing: CHVC and a Magellanic Stream Cloud

This would reveal the fraction of X-ray emission beyond these clouds

Help determine the hot gas component of the Local Group

JNB, Birgit Otte, J. Irwin, M. Putman, E. Lloyd-Davies, C. Breuns (2007)

We see a brightening, not a shadow toward both clouds.

Clouds are interacting with a hot medium within 100 kpc of the MW.

Density/Temp of the hot medium is model-dependent.

Mini-Summary

• There is hot gas around the Milky Way and in the Local Group– Masses are not large (0.1-10E9 Msun)– Cooling times are long (3E8-1E10 yr)

• Observed Lx is consistent with only 0.05 Msun/yr of accretion onto MW (and other spirals)

• HVC probably not dominated by condensations from the dilute hot gas

When the Demons Visit at Night• Where did the Baryons go?

– Cosmological value is 17%– Rich Clusters have not lost their baryons– Spirals (like MW) are missing 2/3 of baryons– Galaxy groups (T < 1 keV) also missing most of their baryons within r1000

• Good News – Bad News– Lots of gas available for accretion– … but it’s nowhere in sight (unbound)

• Make the galaxy, then blow out the baryons in a “superwind” (this must occur, but when?)– Pollutes the surroundings– Need to drive the gas way away from galaxy (to the outer parts of

the groups)

• Other Possibility: The Gas Never Fell In– Entropy floor (preheating is 0.4 keV; 5E6K)

• Need about 1 SNe per 500 Msun of gas

– Enrich the metals by distributed SNe (Pop III)• 0.2 Solar metals is also 1 SNe per 500 Msun of gas

– Naturally solves the G-dwarf problem (accrete enriched gas to make the disk)

– Need to form some parts of galaxy before preheating (halo stars; dwarf galaxies)

– Predict that this SNe heating occurs at z > 2 (Pop III ?)• Disk is 10 Gyr old

– Has a significant effect on modeling (e.g., Davé): gas supply is hot

• How are galaxies so smart?– Tight relationship as more baryons are lost– Clever feedback/formation scheme?

5E6 K

Poor Groups