Galaxies Astro 530 Prof. Jeff Kenney. CLASS 16 March 28, 2018 Molecular Gas in Galaxies & Measuring Star FormaHon Rates in Galaxies

Transcription

1 Galaxies Astro 530 Prof. Jeff Kenney CLASS 16 March 28, 2018 Molecular Gas in Galaxies & Measuring Star FormaHon Rates in Galaxies 1

2 when to make up class? Thurs 1:00? 2

3 Importance of Molecular gas 1.) it s the medium in which stars form 2.) H 2 dominates the ISM mass in centers of gas- rich galaxies; may be dominant form of mass in centers of some galaxies (exceeding mass in stars and dark mazer) 3

4 relahons between CO emission and H 2 mass N(H 2 ) = X I CO = (X/Ω b ) T b (v,ω) dω dv (K km/s) doesn t work for indiv lines- of- sight, only works if averaged over a whole cloud or many clouds X ~ 2x10 20 <n/200 cm - 3 > 1/2 <T/6K> H 2 cm - 2 (K km/s) - 1 X 2.0x10 20 M sun Mpc Jy km/s M(H 2 ) = 8.0x10 3 D 2 S CO (v) dv these equations valid if: a. clouds close to Virial Equilibrium b. cloud properties similar to local Milky Way clouds c. Z>0.5 Z sun [ if Z<0.5Z sun then d CO < d H2 ] 4

5 classical size- linewidth relahon for molecular clouds Size- linewidth relahon for molecular clouds in nearby Milky Way σ~d 0.5 (Larson 1981) Indicates turbulence important throughout molecular ISM Also part of evidence that many molecular clouds are self- gravitahng and in virial equilibrium Size- linewidth relahon for 230 molecular clouds In the Milky Way (Solomon etal 1987) (mid-galaxy 3<R<10 kpc) 5

6 modern size- linewidth relahon for Milky Way molecular clouds shows more variahon in cloud properhes inner galaxy R< 1kpc (gold) outer galaxy R>10 kpc (black) mid-galaxy 3<R<10 kpc (blue, red) Heyer & Dame x more KE than virialized cloud virial equilibrium σ v 2 ~ Σr size- linewidth relahon not uniform throughout galaxy (Larson s Law not obeyed by all clouds!) some clouds not virialized 6

7 if clouds obey virial theorem: <v 2 > = GM/r σ 2 v = <v 2 > ~ GM/r ~ GΣr 2 /r ~ GΣr since Σ = M/r 2 7

8 Molecular Clouds in outer Milky Way have weaker CO emission part of this is a lower mass surface densities of clouds, but part might also be a different CO-H 2 relation Heyer & Dame

9 CO- H 2 relahonship not universal! X factor depends on n 1/2 /T in best(?) case of ophcally thick CO over whole H 2 - dominated region of self- gravitahng gas cloud At low metallicity (Z<~0.5Z sun ), ophcally thick CO covers only small part of H 2 - dominated region of gas cloud (CO only in cores ) à CO not good H 2 tracer in low mass galaxies Not all molecular gas in self- gravitahng clouds à some diffuse molecular gas with high L(CO)/M(H 2 ) Cloud shadowing: clouds which overlap in space & velocity will be hidden 9

10 Main molecular lines observed in other galaxies CO = 12 CO is brightest line, traces gas with densihes ~10 3 cm - 3 CO(1-0) lowest transihon, easiest to excite at low temperatures CO(2-1), CO(3-2), CO(4-3), etc. higher rotahonal transihons, trace hozer gas; rahos give informahon on gas temperature 13 CO and C 18 O more ophcally thin than 12 CO; gives informahon on ophcal thickness (& therefore on diffuse gas) HCN, HCO+ (10 5 cm - 3 ); CS (10 4 cm - 3 ) trace gas with higher densihes; together with CO gives informahon on gas density 10

11 other way to eshmate H 2 masses from molecular line observahons observe ophcally thin species (e.g. C 18 O) in several different transihons determine column density of that molecule along the line- of- sight, e.g N(C 18 O) adopt your best guess for the N(C 18 O)/N(H 2 ) raho to get N(H 2 ) 11

12 CO map of Milky Way Magellanic Clouds Orion Heyer & Dame

13 B Unsharp Masked (shows Dust ExHncHon) NGC 891 CO WIYN images (Howk & Savage 1999) VerHcal distribuhons of CO in spirals NGC 891 is edge- on Sb similar to Milky Way VerHcal scale heights MW(sol neigh) N891 (R>1 kpc) CO 50 pc 100 pc OB * s 50 pc FG * s few 100 pc Molecular gas is thinnest disk component. Newly formed stars have same scale height as molecular clouds in which they form Scoville etal 1993 As stars age, they heat up and increase their scale height Molecular gas scale height likely smaller near galaxy centers (thinner disks near center) 13

14 CO optical (stars) Radial distribuhons of CO vs. stars Similar radial distribuhons on average, but more structure in CO than stars In central ~1-2 kpc (or bulge region), some galaxies have lots of molecular gas, others have relahvely lizle Regan etal 2001 BIMA SONG 14

15 Radial distribuhon of H 2 in Milky Way (from R=3-16 kpc) for R<8 kpc, M(H 2 )~M(HI) for R>8 kpc, M(H 2 )<<M(HI) Heyer & Dame

16 Milky Way radial gas distribuhons circumnuclear gas concentration in central ~1 kpc bar region new calibration of CO: H 2 should be ~3 x lower! Scoville 2012

17 HI vs H 2 radial distribuhons in spirals H 2 fills in central plateau or hole in HI in 4 of these spirals; central depression in both HI & H 2 in one of these spirals Kenney & Young 1989 Radius (arcmin) 17

18 CO vs. other radial distribuhons in spirals H 2 Log Gas Mass Density in M sun pc - 2 or Intensity Starlight (B) HI Sb spiral NGC 4192 Kenney & Young 1989 Radius (arcmin) CO radial distribuhons most similar to tracers of massive star formahon (i.e. Ha, FIR, radio conhnuum) CO radial distribuhon similar to disk starlight in many cases - > evidence that stellar disk is exponenhal because it formed from exponenhal gas disk (dissipahon & angular momentum transport?) Central HI hole or depression or plateau filled in with molecular gas (H 2 ) in many (but not all) galaxies Some galaxies have relahve lack of CO in center, roughly coincident with bulge

19 Stellar disks are thin and exponenhal because they form from gas disks that are thin and ~exponenhal, due (in part) to gas dissipahon. Stellar bulges are fat and r 1/4 because they form from star systems that reach equilibrium without dissipahon. 19

20 CO mapping of grand design spiral M51 BIMA SONG Helfer etal 2003 Combined interferometer & single dish data à get good resoluhon & include all flux 20

21 CO on R Spiral structure in CO & other tracers in M51 CO on Ha Young stars (HII regions) located downstream from CO peak in some parts of spiral arm CO on Radio cont (non- thermal) CO on Radio Cont (thermal) indicates Hme sequence for star formahon Garcia- Burillo etal 1993 Large arm- interarm contrast (~3:1) in CO: many molecular clouds form in spiral arms 21

22 Galaxy centers: Gas responds strongly to stellar bars M101 M101 has weak stellar bar Molecular gas has strong response, piles up along leading edge of bar GravitaHonal torques can drive gas inwards 22

23 CO in strongly Barred Galaxies CO and Ha gas kinemahcs Laine etal 1999 NGC 7479 BVR image CO peaks lies on dust lane at leading edge of stellar bar Large velocity jump ( km/s) across the bar!! Dense gas piles up due to shock at leading edge of bar 23

24 Circumnuclear Star FormaHon in Spiral Galaxies NGC 3351 M95 D=10 Mpc Star- forming ring deep inside bar at r=6 =300pc in NGC 3351 BVR Herschel Telescope Knapen 24

25 Very Different DistribuHons of Gas and Young Stars in non- starburst NGC 3351 CO(1-0) 2 OVRO Jogee+05 K HST UV on Hα CO Star- forming ring R=5 =300 pc molecular gas exists both inside and outside of star forming ring & within the ring is concentrated at 2 ends 25

26 Molecular Gas ProperHes of Galaxies How much molecular gas? In which ISM phase is most of the gas mass? 26

27 M(H 2 ) vs. M(stellar) in massive nearby galaxies M(H 2 ) = M star M(H 2 ) = 0.01 M star literature compilahon: much larger range than careful dedicated survey suggests large errors for some M(H 2 ) values in literature COLD GASS Saintonge etal 2011 COLD GASS CO(1-0) survey 350 nearby galaxies with M(star)>10 10 M sun All morphological types Cold gas mass frachon ranges from M(H 2 )/M(star) <0.01 to 0.20 Average for detected galaxies

28 M(H 2 )/M star vs. key properhes in massive nearby galaxies M(H 2 )/M(star) drops for: M(star)>10 11 M sun µ(star)> M sun kpc - 2 R90/R50>2.6 E s, large bulge S0 s and Sa s have rela>vely li@le cold gas CorrelaHon between M(H 2 )/M(star) and NUV- r higher cold gas frac>ons in bluer galaxies (which have higher SFRs) COLD GASS 28 Saintonge etal 2011

29 M(H 2 ) vs. M(HI) in massive nearby galaxies M(H 2 )/M(HI) varies by 2(+) orders of magnitude in massive galaxies, Range: M(H 2 )/M(HI) <0.03 to >3 trends with stellar mass, color, etc are very weak (so hard to predict ) COLD GASS Saintonge etal

30 Molecular Gas ProperHes of Galaxies How much molecular gas? In which ISM phase is most of the gas mass? M(H 2 )/M(HI) ~ for nearby spirals, less for many E s and S0 s M(HI+H 2 )/M(stars) ~ for nearby spirals (generally higher in galaxy centers, low mass late type galaxies, mergers, young galaxies; generally lower for E s and S0 s) L(CO)/M(HI) and L(CO)/M(stars) lower in low- mass (dwarf) galaxies May be due to different X (lower metallicity) or lower M(H 2 )/M(HI) - - it is probably both! 30

31

32 one of main ways that galaxies evolve is by converhng gas into stars à the star forma+on rate (SFR) is one of the most important measures of galaxy evolu+on How can we measure the SFR in galaxies?

33 How can we measure the rate of star formahon in galaxies? main approach: observe radiahon from stars that must be young (it is hard to eshmate ages of stars!) use massive stars: massive stars don t live long, so every one must be young massive stars are hot & luminous, so they produce lots of UV radia>on

34 main sequence life>me of star τ MS τ MS (M) ~ (M/M sun ) c τ sun τ MS (M) ~ (M/M sun ) c yrs

35 main sequence life>me of star τ MS τ MS (M) ~ (M/M sun ) τ sun τ MS (M) ~ (M/M sun ) yrs this is approximahon which is a bit off for the most massive and least massive stars lifehme of 10 M sun star is 10 7 years

36 Lyman conhnuum UV photons (E>13.6 ev, λ< µm) ionize H atoms Only blackbodies hozer than T>20,000K emit significant quanhhes of Lyman conhnuum UV photons OB stars have T>20,000K Main sequence (MS) stars have large luminosity White dwarfs hot enough but small size means small luminosity MS OB stars don t live long so any MS OB star is young & therefore formed recently

37 CounHng Lyman conhnuum photons to eshmate massive star formahon rates Basic idea: Lyman conhnuum photons (extreme ultraviolet photons, λ<91.2nm) in galaxies are produced (mostly) by hot young massive stars which live for such a short Hme that every one we see must have formed recently

38 Lyman con>nuum photon rate. N Ly (tot) = N(>10 M sun ) 10 Mu N(M) N ly (M) τ MS (M) dm Mu N(M) dm where: 10 N Ly (tot) = total # Lyman continuum photons from galaxy. N(>10 M sun ) = stellar birthrate (stars born per year) for stars with M>10 M sun (only ones that emit LyC photons) M 2 N(M) = Ψ(M) dm #stars formed with masses between M 1 and M 2 M 1 Ψ(M) = dn/dm = IMF = initial mass function (# stars formed per unit mass interval) N ly (M) = Lyman continuum photon rate for star of mass M τ MS (M) = main sequence lifetime of star M u = upper mass cutoff (most massive stars which form) what we want!!

39 Lyman con>nuum photon rate relahon for Lyman conhnuum photon rate assumes that there are enough individual massive stars at any one Hme that we can treat problem sta+s+cally by weighing the contribuhon from each star by its lifehme and abundance relahve to stars of different masses this is OK for most late type galaxies which have 100 s s of massive stars, but approach breaks down for smaller numbers (& smaller regions within galaxies). also breaks down if SFR (N) has not been constant for last ~10 7 years

40 Lyman con>nuum photon rate per star N ly (M) = Lyman conhnuum photon rate for star of mass M N ly (M) = 3.8x10 42 (M/M sun ) 3.86 N ly (M) = 2.4x10 35 (M/M sun ) 8.67 (only appx, Panagia 1973) M>32 M sun 9 M sun >M>32 M sun N ly (M) = 2x10 29 (M/M sun ) log(M/Ms) M>10 M sun (Tutokov & Krugel 1981)

41 main sequence life>me of star τ MS τ MS (M) ~ (M/M sun ) τ sun τ MS (M) ~ (M/M sun ) yrs this is approximahon which is a bit off for the most massive and least massive stars lifehme of 10 M sun star is 10 7 years

42 ini>al mass func>on IMF Ψ(M) N(M) = Ψ(M) dm #stars formed with masses between M 1 & M 2 M 1 M 2 Ψ(M) = dn/dm = IMF = inihal mass funchon (one way to define) dn dm = Ψ(M) ~ M -α dn ~ M dlogm -γ α = γ + 1 one way to define IMF another way to define IMF Salpeter (1955) α = historical Kroupa & Weidner (2003) α = for M sun α = -1.3 for M sun modern values Chabrier (2003) similar result but more complex form

43 Lyman con>nuum photon rate. N Ly (tot) = N(>10 M sun ) 10 Mu N(M) N ly (M) τ MS (M) dm Mu N(M) dm where: 10 N Ly (tot) = total # Lyman continuum photons from galaxy. N(>10 M sun ) = stellar birthrate (stars born per year) for stars with M>10 M sun (only ones that emit LyC photons) M 2 N(M) = Ψ(M) dm #stars formed with masses between M 1 and M 2 M 1 Ψ(M) = dn/dm = IMF = initial mass function (# stars formed per unit mass interval) N ly (M) = Lyman continuum photon rate for star of mass M τ MS (M) = main sequence lifetime of star M u = upper mass cutoff (most massive stars which form) what we want!!

44 relahng LyC rate to birthrate to MSFR to SFR relate SFR (in M sun yr - 1 ) to birthrate (in #stars yr - 1 ). SFR = N <M> where <M> = M U M L M L M U N(M) M dm N(M) dm MSFR (M>10 M sun ) = 1.2 x N Ly (tot) (M sun yr -1 ) (photons sec -1 ) Kennicutt & Evans 2012 assumes: Kroupa IMF, more accurate form of N Ly (M), solar metallicity for all stars (N ly ~ Z -0.4 ); this is average over 10 7 yrs, lifetime of 10 M sun star SFR(tot) = constant x MSFR (M>10 M sun )

45 relahng LyC rate to birthrate to MSFR to SFR relate SFR (in M sun yr - 1 ) to birthrate (in #stars yr - 1 ). SFR = N <M> where <M> = M U M L M L M U N(M) M dm N(M) dm MSFR (M>10 M sun ) = 1.2 x N Ly (tot) (M sun yr -1 ) (photons sec -1 ) Kennicutt & Evans 2012 assumes: Kroupa IMF, more accurate form of N Ly (M), solar metallicity for all stars (N ly ~ Z -0.4 ); this is average over 10 7 yrs, lifetime of 10 M sun star SFR(tot) = 6.3 MSFR (M>10 M sun ) for Kroupa IMF

46 most of stellar mass is in low mass stars! SFR(tot) = 6.3 MSFR (M>10 M sun ) à most of gas mass consumed by star formahon goes into low mass stars which are not easily observable! à correchon factor to get from MSFR to SFR is large and depends on IMF, which is very hard to measure in other galaxies

47 We can t directly measure most of the Lyman conhnuum photons produced in absorption cross section 13.6 ev = 912 nm Lyman conhnuum photons (~50nm<λ<91.2nm) are those which easily ionize H atoms (absorphon cross- sechon ~ λ 3 ). There are so many H atoms in space & the absorphon cross- sechon is so large that these photons rarely escape the galaxy.

48 EsHmaHng SFRs in Galaxies I. Lyman UV conhnuum (λ<91.2nm) A. recombinahon lines (e.g. Hα) B. thermal radio conhnuum II. UV conhnuum (91.2nm <λ<280nm) III. Far- Infrared conhnuum IV. Non- thermal radio conhnuum MulH- wavelength observahons combine different tracers to achieve greater accuracy