Galaxies Astro 530 Prof. Jeff Kenney CLASS 20 April 9, 2018 Hot Gas in Galaxies 1
Reading week Mon Apr 30 no lecture (I compressed schedule); last lecture is Wed Apr 25 Homework HW 9 due Mon Apr 16 HW 10 due Mon Apr 23 HW 11 due Mon Apr 30
galaxy talks (reading week): 11 talks @ 15+5 min/talk 2 days @ 1.5 hr/day + 1 days @ 1 hr/day tenta%ve schedule: Mon Apr 30 4-5:30pm (4 talks) Dwarfs+Bars (yotam, joel, dhruba, sarah) Tues May 1 4-5:30pm (4 talks) Starbursts+RPS (lily, malena, ^m, michael) Wed May 2 4-5:00pm (3 talks) AGN+clusters (uddipan, daniel, pra^k)
final exam proposed exam ^me is Mon May 7 at 9am. we don t have to have it then. finals week: Fri May 4 Wed May 9
MAJOR PHASES OF ISM IN MILKY WAY * PHASE OF ISM T(K) n (cm - 3 ) h (pc) f vol f mass Hot ionized medium (HIM) >10 6 10-3 3000 0.5 0.04 Warm ionized medium (WIM) 8000 0.3 900 0.2 0.10 Warm neutral medium (WNM) 6000 0.3 400 0.3 0.3 Cold neutral medium (CNM) 100 20 140 0.02 0.2 Molecular medium (MM) 10 10 2-10 3 70 0.001 0.3 T = temperature N = volume density h = ver^cal scale height f vol = volume filling factor f mass = ISM mass frac^on * Hot gas in galaxies has temperature range 10 6 10 8 K Values given are representa^ve values for Milky Way. h and f vol are values within a few kpc of solar neighborhood; the inner ~kpc and outer galaxy are different. Ref: Brinks 1990; Boulanger & Cox 1990; Kulkarni & Heiles 1988
MAJOR PHASES OF ISM IN MILKY WAY * PHASE OF ISM descriphon Traced by Hot ionized medium (HIM) Warm ionized medium (WIM) Warm neutral medium (WNM) Cold neutral medium (CNM) Coronal gas produced by supernovae HII regions around massive stars & throughout ISM Diffuse clouds & envelopes around molecular clouds Dense sheets & filaments; envelopes around molecular clouds X- Ray con^nuum emission Hα+other recombina^on lines; radio con^nuum emission HI in emission HI in absorp^on Molecular medium (MM) Cold, dense, gravita^onally bound clouds CO and other emission lines Ref: Brinks 1990; Boulanger & Cox 1990; Kulkarni & Heiles 1988
M31 Spiral M31 in X- Rays CHANDRA red 0.3-1 kev green 1-2 kev blue 2-8 kev 3 main sources of X- Ray emission in galaxies: Hot Gas (T~10 7 K) - - bremsstrahlung - - extended, sor spectrum Binary Stars (with accre^ng NS or BHs) - - point sources, hard spectrum AGN nuclear point source, hard spectrum (not all galaxies have AGN)
M31 in X- Rays color coded by X- Ray temperature red = cool yellow = s^ll cool green = warm blue = hot diffuse emission from hot gas is cooler than most compact sources hot gas extent r=2 kpc ~ bulge- dominated region 1.5 kpc most compact sources are stellar sources in M31; some are background AGN 28 ~ 6 kpc per side
X- Ray & 24µm IR in M31 red: Spitzer's 24µm green: low- energy X- rays from Chandra (0.5-2.0 kev) blue: high- energy X- rays from Chandra (2.0-4.0 kev). Most dust & SF beyond region of hot gas, but some cold ISM in central 2 kpc. Ra^o of hot gas to cold gas drops with increasing radius. Hot gas dominates in bulge- dominated region, cold gas dominates in disk- dominated regions but some of both everywhere. 28 ~ 6 kpc per side less X- Rays where there are dust lanes - - absorphon of x- rays by neutral gas
Sombrero Galaxy M104 large bulge Sa X- Ray op^cal IR 24µm hot gas has more spheroidal distribu^on, like stellar bulge cold ISM has more disky distribu^on, like stellar disk
Chandra X- Ray Spectra of E s High mass E s: 3-15x10^10 Lsun Hot gas dominates 2 components Low Mass E s: 0.5-3x10^10 Lsun (and spirals) Binaries Dominate (bremsstrahlung) Humphrey & Buote 2006
X- Ray vs Op^cal luminosity for Ellip^cals different behavior at high and low mass at high L B L x ~L B 2 (hot gas dominates) at low L B L x ~L B (stellar sources dominate) expected contribu^on from hot gas ** large scatter not well understood! expected contribution from stellar sources, L x ~L B Matthews 2003
why does L X ~ L B 2 for big E s? higher mass galaxies have their gas heated to higher temperatures stellar velocity dispersion and expected gas temperature both higher in more massive galaxies so massive galaxies have more gas mass and higher gas temps (why rela^on is not linear)
analyze the X- Ray surface brightness radial profile for an ellip^cal we can learn the amount of hot gas & the dynamical mass of the galaxy!
X- Ray surface brightness radial profile of massive ellip^cal & inferred gas density profile I X (R) Einstein n e (r) NGC 4472 X-Rays τ cool (r) Chandra projected radius R (arcsec) 3D radius r (arcsec) Trinchieri+1986
azimuthally averaged surface brightness profile I X (R) = I X (0) [ 1 + ( ) 2 ] - 3β+1/2 R a x R = projected radius a x = core radius ~ 2 kpc typically b = index ~ ½ typically, so R I X (R) ~ [ 1 + ( ) 2 ] - 1 a x similar to King & Hubble op^cal light profiles I X (R) projected radius R (arcsec)
what is producing the X- Ray emission? I ν = ε ν (l) e -τ dl = ε ν (l) dl if op^cally thin τ=0 ε ν = volume emissivity for free- free emission (thermal bremsstrahlung): ε ν ~ Z 2 n p 2 T - 1/2 e - hν/kt NOTE: emissivity goes as density SQUARED
surface brightness can be related to gas density I X (R) = I X (0) [ 1 + ( ) 2 ] - 3β+1/2 R a x - l for free- free emission, if gas is isothermal and has a spherically symmetric gas distribu^on: n p (r) = n p (0) [ 1 + ( ) 2 ] - 3β/2 r a x R r l r = galactocentric (3D) radius +l
how much hot gas? given gas density distribu^on we can calculate M gas M gas = ρ gas dv = m p n p dv m p = proton mass M gas ~ r 3/2 which diverges.. so n p (r) profile cannot continue to very large r M gas /M star = 0.001-0.1 for E s M halo <10 13 M sun M gas /M star = 1-5 for clusters M halo >10 14 M sun out to where X-Rays detected more for larger galaxies most galaxies <0.01 so gas mass much less than stellar mass gas dominates stars on scales of clusters and larger!
using hot gas in E s and clusters to measure total masses X- Ray emission from hot gas may be used to measure the gravita^onal poten^al IF gas is close to hydrostaic equilibrium Euler equa^on in case of no mass flow: F - P = 0 F = force per unit volume when outward gas pressure balanced by inward force of gravity, in spherically symmetric system: this is equa^on of hydrosta^c equilibrium for spherical gas mass
using hot gas in E s and clusters to measure masses for ideal gas law P = k ρ T µ m p combining ideal gas law and hydrosta^c equilibrium, one can show: kt(r) d log ρ d log T Gµm p d log r d log r M(r) = - ( ) ( + ) r T(r) d log ρ d log T r 10 7 K d log r d log r 100kpc M(r) = - 3x10 12 ( )( + )( ) M sun galaxy mass can be measured, if radial gradients in T and r can be measured, and gas close to hydrosta%c equilibrium
Measure T(R) and ρ(r) from X- Ray observa^ons Gas Temperature vs. radius Chandra X- Ray maps of E s From X- Ray surface brightness profile I x (R) and gas temperature profile T(R) (determined from spectrum), can determine gas density profile n p (r) Humphrey etal 2006 Gas density vs. radius
M/L K vs. radius (R e ) Ellip^cal mass distribu^ons Enclosed Mass vs. radius stars gas DM Humphrey etal 2006 X- Rays from hot gas allow dynamical mass to be measured to large radii in E s M/L is low inside the op^cal radius R<R e where mass is dominated by stars M/L rises by factor of ~10 out to ~10R e, indica^ng lots of dark ma~er R>R e massive E s have M tot ~ 10 13 M sun within r~100 kpc
but is gas really smoothly distributed? is it really in hydrosta^c equilibrium? we will return to this later!
things we d like to understand about hot gas in galaxies origin of gas hea^ng & cooling of gas physical state of gas (e.g., is gas in pressure equilibrium?) - > spa^al distribu^on, substructure of gas, elemental abundances, help us know these things
Origin of hot gas in E s 1. internal - - stellar mass loss 2. external accre%on (either primordial from cosmic web, or fallback from early intense starburst oulow) High metallicity of gas (solar to slightly supersolar in massive E s) indicates stellar mass loss or starburst fallback important Abundance ra^os imply ~2/3 of Fe in ISM produced by Type Ia supernovae (similar to MW)
Recycling by stars back into the ISM Low mass stars: (M<8Msun): Planetary nebulae Medium mass stars in binaries: Type Ia Supernovae High mass stars (M>8Msun): Type II Supernovae
Stellar mass loss in galaxies Most of mass injected into ISM by red giants and planetary nebulae (most of energy injected by high mass stars). Since these star contribute much of galaxy luminosity, their death rate is propor^onal to op^cal luminosity:. M * /L B = 1.5x10-11 M sun yr - 1 L sun - 1 for old stellar populahon few- 10x higher for younger stellar popula^ons
Stellar mass loss in galaxies Most of mass injected into ISM by red giants and planetary nebulae (most of energy injected by high mass stars). Since these star contribute much of galaxy luminosity, their death rate is propor^onal to op^cal luminosity:. M * /L B = 1.5x10-11 M sun yr - 1 L sun - 1 for old stellar populahon few- 10x higher for younger stellar popula^ons rate of stellar mass loss M * in old stellar popula^on determined mostly by: 1. Difference between ini^al and final masses of stars currently undergoing evolu^on off main sequence 2. Death rate of such stars
. Stellar mass loss in galaxies M * /L B = 1.5x10-11 M sun yr - 1 L sun - 1 for old stellar populahon few- 10x higher for younger stellar popula^ons Over life^me of 10 10 yrs, E with L~10 11 L sun should have M gas >10 10 M sun injected into ISM from stellar mass loss Consistent with gas mass of brightest Es, but smaller E s have less gas than rela^on predicts Suggests that in smaller E s gas is lost from galaxy, or cools to lower temps (either might depend on galaxy mass)
Hea^ng of Gas Large gas masses with T>10 7 K implies lot of energy. What is source of energy? Not well understood. Unse~led ques^on but related to inhibi^ng SF in most massive galaxies, and stopping the growth of most massive galaxies Possibili%es: 1. stellar mo^ons 2. supernovae 3. AGN 4. gravita^onal
Hea^ng of Gas 1. stellar mohons gas- losing stars are moving at large random veloci^es, so gas will be injected with same KE per unit mass as stars; this injected gas will collide with other gas thermalize at temperature: T = µm p σ *2 /k = 7x10 6 K (σ * /300 km s - 1 ) 2 Since T obs > 10 7 K, stellar mo^ons important but might not be dominant hea^ng source
2. Supernovae (Type Ia) Hea^ng of Gas Gas hea^ng rate depends on SN frequency and coupling of SN energy to hot gas. Could provide more than enough energy to heat gas, but coupling of SN energy to gas is uncertain.
Giant HII region NGC 604 in M33: energy from massive stars & SN heats gas to 10 7 K composite hot gas fills in space between Hα filaments On W (right) side, + hot gas mass & luminosity consistent with hea^ng from colliding stellar winds from 200 massive stars (earlier evolu^onary stage) Op^cal Hα HST Chandra sor X- Ray On E (ler) side, higher hot gas mass & fewer massive stars suggest hea^ng by supernovae and massive stars in past (later evolu^onary stage) Tuellmann et al 2008
Spiral galaxy M83 in X- Rays M83 in X- Rays Chandra red: low energy green: medium energy blue: high energy Long etal 2012 hot gas in star- forming regions: energy from massive stars & SN clearly heats lots of gas to 10 7 K
Antennae (major merger) in X- Rays, IR, & Op^cal composite X- Rays Chandra Op^cal HST IR Spitzer IR emission (24 µm?) shows regions of intense star forma^on Diffuse X- Ray emission from T~10 7 K gas surrounds regions of ac^ve star forma^on Note lack of x- rays where dust lanes are strong. This is due to absorp^on of sor x- rays by neutral gas associated with the dust.
How do we figure out the Energy deposi^on rate into ISM from SN?
Energy deposi^on rate into ISM from SN. E ~ η coup E SN SNR E SN energy per supernova (erg) SNR supernova rate of galaxy (# SN/yr) η coup coupling efficiency (dimensionless) very uncertain
Hea^ng of Gas 3. AGN X- Ray 0.1-6 kev Ellip^cal M84 in Virgo Cluster Radio 5 GHz M84 E in Virgo Cluster X- Ray and Radio Substructure associated with radio AGN features clearly show AGN affects hot gas in many galaxies. Could provide more than enough energy to heat gas, but coupling of AGN energy to hot gas is uncertain. Significant hea^ng source in center of galaxy, where gas is densest & predicted cooling ^mes are shortest. AGN not always on. Are they on oren enough to maintain hot ISM?
Hea^ng of Gas 4. GravitaHonal KE of accre^ng gas may be converted to thermal energy of gas. Is accre^on rate high enough to provide necessary hea^ng?
but is gas really smoothly distributed? is it really in hydrosta^c equilibrium?
Is gas really in hydrosta%c equilibrium? Hot gas distribuhons in elliphcals can be smooth or have lots of small- scale or large- scale structure compare these 2 large ellip%cals in the Virgo Cluster Chandra X- Ray on HST op^cal Chandra X- Ray M60 (NGC 4649) gas distribu^on smooth no recent AGN disturbance M84 (NGC 4374) gas has large scale structure recent AGN disturbance
X- Ray op^cal Substructure of Hot Gas in Ellip^cals sample selected to be most regular, least affected by AGN Statler & Diehl 2005 Structure in X- Ray maps à hot gas is not fully in equilibrium - - probably due to energy injec^on by AGN & supernovae Clumpiness of X- Ray emission à gas densi^es derived assuming of smooth distribu^on will overes^mate true gas densi^es
examine assump^on of hydrosta^c equilibrium in simple case: ellip^cal with smooth gas distribu^on Chandra X- Ray on HST op^cal X- Ray distribu^on very smooth no recent AGN disturbance M60 (NGC 4649) one of largest E s in Virgo Cluster SN disturbances seem modest
gravita^onal poten^al of central quiescent BH increases temperature of central bin? T(r) M60 ρ(r) Humphrey+2013 no significant substructure detected in Chandra X- Ray map so maybe assump^on of hydrosta^c equilibrium is OK in M60
how good are mass es^mates from assuming gas in hydrosta^c equilibrium? circular speed (km/s) M60 stellar dynamics hot gas in hydrosta^c equilibrium Humphrey+2013 radius (kpc) P = P therm + P turb + P mag + P cr circular speed curve from stellar dynamics is ~10% higher than that from assuming hydrosta^c equilibrium for hot gas suggests addi^onal support for hot gas beyond thermal pressure support from subsonic gas mo^ons (either bulk or turbulent), or magne^c fields or cosmic rays probably important even in simple case of E with a very smooth X- Ray distribu^on (M60), ~25% of pressure is non- thermal (turbulent gas mo^ons?) similar to clusters
Is gas really in hydrosta^c equilibrium? NO, NOT COMPLETELY Sub- structure in X- Ray distribu^ons in some E s indicates gas distribu^on is not homogenous & smooth & can t be in pure hydrosta^c equilibrium Cooling ^me of gas very short in central regions; AGN disturbances can be large in centers
Is gas really in hydrosta^c equilibrium? NO, NOT COMPLETELY, BUT Humphrey etal (2006) sample selected to be those E s with less X- Ray substructure Humphrey etal (2006) derived mass profiles are smooth & sensible Can include addi^onal term(s) to account for non- thermal pressure support from random gas mo^ons & magne^c fields in equa^on of hydrosta^c equilibrium Cooling ^mes much longer in outer galaxies, so hydrosta^c equilibrium may not be bad appx for outer galaxies If E s didn t contain large dark ma~er haloes, they could not bind 10 7 K gas! BOTTOM LINE: Despite complexi^es in hot gas distribu^ons in some E s, gas close enough to hydrosta^c equilibrium in most E s to have it work as decent tracer of total mass profile, especially if non- thermal pressure components can be es^mated
rapid cooling predicted in centers of E s and clusters Cooling rate of gas ~ n e n p ~ n p 2 photon energy escapes the galaxy, cooling the gas (photon luminosity = cooling rate) cooling ^me is short (~10 7 yr) in central parts of galaxies, where gas density is high, and longer (~10 10 yrs) in outer parts, where gas density is low Unless gas is somehow reheated, gas in centers will rapidly cool to T<<10 7 K n e (r) t cool (r) cooling ^me (years) 3D radius r (arcsec)
many E s and clusters have cool cores, due in part to more rapid cooling in centers NGC 4472 (M49) X- Ray Chandra much cooler in center (shortest cooling ^me) more constant outside center (close to isothermal)
Cooling Flows in E s (?) Cooling rate of gas ~ n e n p ~ n p 2 cooling cooling ^me is short (~10 7 yr) in central parts of galaxies, where gas density is high, and longer (~10 10 yrs) in outer parts, where gas density is low à Unless gas is somehow reheated, gas in centers will rapidly cool to T<<10 7 K flows as it cools, ρ increases so as to maintain constant P; volume decreases and gas element moves (slightly) radially inward Suggests that cooling flows could exist
Do Cooling Flows exist? Cooling? There is definitely cooling gas (it produces the X- Ray emission we observe), but the gas could be reheated, so not clear there is net cooling. Flows? There may be modest inward mo^on, but not much, and the main ques^on is about the cooling.
Do Cooling Flows exist? There is evidence of some cooling gas in some E s (and clusters of galaxies) e.g. filaments of T~10 4 K gas (producing Hα emission), but not very much. Cooling rates of M dot ~ (2µm p /5kT) L x are predicted if gas not reheated, which corresponds to ~1000 M sun yr - 1 in some cd galaxies! X- Ray spectroscopy with XMM shows li~le or no gas with T~ few x 10 6 K, which should be present if there were net cooling, so gas is not cooling very much. Gas is probably reheated somehow not well understood but good possibility is AGN which deposits energy in center just where predicted cooling rates are maximum
Cluster MS0735 Radio jets from AGN push out hot gas, crea^ng cavi^es in X- Ray halo B&W op^cal HST Blue X- Ray Chandra Red Radio VLA McNamara etal 2005
galaxy cluster Abell 2597: cool core cluster op^cal brightest cluster galaxy X- Ray Hα radio on X- Ray high frequency radio jets from recent AGN outburst only in very center ghost cavi^es in X- Ray map probably filled by low frequency radio emission from older AGN outburst AGN par^ally heats gas but not enough to prevent some gas from cooling & forming stars (Hα from HII regions and shock- excited filaments) Voit+2015
a feedback loop of cooling and hea^ng, s^fling star forma^on in the middle of these galaxy clusters? galaxy cluster Abell 2597 X- Ray, Hα & op^cal Voit+2015 AGN outbursts heat gas and can prevent cooling for a while, but then gas cools some, the central BH accretes un^l another AGN outburst happens