Star Formation at Low Metallicity ADAM LEROY (NRAO) 1 Things should be different at low metallicity. 2 Scaling relations: SFR, CO, HI, and stars. 3 PDR Structure: H2, HI, CII, and CO 4 Molecular clouds at low metallicity. 5 What to look for in the next few years… 1 Less metals mean less dust. ! ! "#$%! &%'! !"#$%$&%'(#) *($+&) &,) '(-(.+/#) 0&12(*/3) $&) 1/$(--+0+$45! ()*+,! -+,+./! 0+12,3! 40! 563!F 0+13!ISHER +0! ET 47! AL "4-893!. (SUBM :%! ;.),3! +,0)!GALAMETZ 47*,8<3! 5=)! ET 73+9>/?!AL. (2009), =3,,@A7)=7! DRAINE ,)=! ET AL. (2007) 135+,,4*45/!-+,+.430!;(B:!C2892,3!*49*,3D!+7<!563!EBF:G!C>,83!*49*,3D%!H63!0),4<!,473! 9329303750! +! ,473+9! *)993,+54)7! >35=337! <805@! 5)@-+0! 9+54)! +7<! 135+,,4*45/?! 035! 5)! 1+5*6!563!B4,A/!;+/?!IJK!0*+5539!40!9329303753<!>/!<+063<!,4730%!!")9!93L3937*3?!=3! 06)=!56+5!#!M=!NO!40!*)70405375!=456!+!2)=39!,+=!93,+54)70642JO!47!=64*6!P$Q!∝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`W! (38S37?! FE(?! #BZF! CT3,-481D_! FZU?! (UB!C"9+7*3D_!B]#U!C$391+7/D_!#XU"@#"E#abUUabU]abUH?!(ZXE?!E#EEU!C#5+,/D_!#UF! CE2+47D%! H640! <3S3,)21375! 6+0! >337! 0822)953<! >/! 563! L87<47-! +-37*430! TB^#H! 1 Covariant with stellar, gas, kinematic structure. Spiral CO HI velocity stars star formation Dwarf Galaxies 1 Covariant with stellar, gas, kinematic structure. IC 2574 HI Dwarf Stars CO Surface Density NGC 3184 Spiral Galactocentric Radius LEROY ET AL. (2008) 1 Structural differences affect cloud formation. Lower columns mean less molecules Molecular to Atomic Gas Ra1o LEROY+KINGFISH IN PREP. 1 Big, high SFR low-metal galaxies (esp. at high z) Evolution with redshift Mass as a third parameter MANNUCCI ET AL. (2009, 2010) 1 Molecular gas (CO) very hard to observe. WLM – first CO detection below 12+log O/H ~ 8.0 ELMEGREEN ET AL. (2013) 1 Things should be different at low metallicity. • Low metals mean less dust • Dust affects visibility of gas (CO), H2-HI balance • Z covariant with gas, stellar, kinematic structure • Column correlates with H2-HI balance • SFR as a third parameter / redshift evolution • CO faint, despite pervasive star formation 2 SFR globally tracks HI, with mass dependence. In the local volume SFR tracks HI, tdep~ 1 to 0.1 Hubble times LEE ET AL. (2009) 2 SFR globally tracks stellar content. Doubling time <~ ½ Hubble time, less clear slope than HI LEE ET AL. (2009) 2 SFR/CO increases as metallicity drops Global ratio SFR/CO sharply increases with dropping Z at high/low z SFR-to-CO Ratio SFR-to-CO SFR-to-CO Ratio SFR-to-CO Metallicity (12 + log O/H) SCHRUBA ET AL. (2012), GENZEL ET AL. (2012) 2 SFR still coincident with H2, local scaling Even as ratio changes, SFR and H2 still show similar distribution. ] -2 kpc From -1 CO yr sun M From Dust SFR Surface Density (UV+IR) [ SFR Surface -2 H2 Surface Density [Msun pc ] BOLATTO, LEROY ET AL. (2011), JAMESON, BOLATTO ET AL. (IN PREP.) 2 SFR and CO track in HI dominated regime. More generally, SFR still tracks molecular gas in HI-dominated regions. ] -2 kpc -1 Atomic Gas (21-cm) Molecular Gas (CO 2-1) yr sun M SFR Surface Density (UV+IR) [ SFR Surface Density -2 Gas Surface Density [Msun pc ] SCHRUBA ET AL. (2011) 2 SFR tracks older stars in nearby dwarf galaxies. Dwarfs in color, spirals in gray ] ] -2 -2 kpc kpc HI -1 -1 Stars yr yr sun sun Spirals M M SFR Surface Density (UV+IR) [ SFR Surface Density (UV+IR) [ SFR Surface Density -2 -2 Stellar Surface Density [Msun pc ] HI Surface Density [Msun pc ] HUNTER ET AL. (1998), LEROY ET AL. (2008) 2 Scaling relations: SFR, CO, HI, and stars. • Global SFR tracks HI (with some slope) • Global SFR tracks stellar content • SFR/CO increases with dropping metallicity • SFR still associated with CO where HI dominates • SFR broadly occurs where there are stars 3 PDR Structure: H2, HI, CII, and CO Dust controls UV extinction and physical sizes of regions Self & cross shielded HII HI H2 C+ + e- à C C + OH à CO + H C+ + S à C + S+ C + γ à C+ CO + γ à C + O C+ C CO increasing A v Av~1 Av~2 Tielens & Hollenbach (1985) HII region PDR Dark cloud T~10,000 K T ~ 1,000 to 10 K T ~ 10 K τCO=1 surface 3 CO and H2 surfaces shrink with decreasing Z Dropping metallicity pushes H2 in, CO in even faster H C+ 2 CO Z MALONEY & BLACK (1988), BOLATTO+ (1999), RÖLLIG+ (2006), BOLATTO+ (2013) 3 PDR Structure: H2, HI, CII, and CO Being atomic and not molecular doesn’t prohibit cold gas Physical conditions similar with molecular and atomic cooling Temperature Molecular Cooling Atomic Cooling Density GLOVER & CLARK (2010, 2012) 3 Stars can form before molecules pervade At low enough metallicity, stars may form before molecules dominate Temperature Time (in free fall times) KRUMHOLZ (2012) 3 PDR structure confuses CO observations Observing these effects complicated by our tracers. GLOVER & CLARK (2012) 3 PDR structure confuses CO observations $ !0.4 ' Shielding with constant surface density ~ 1.5 exp& ) (Wolfire et al. 2010) % DGR" #100 ( 25 Fig. 9.— Conversion factor, estimated from dust-based approaches, as a function of gas-phase abundance. (Left) Color points show estimates for very nearby galaxies Israel (1997), Madden et al. (1997, based on [CII]),Leroyetal.(2007),Gratieretal.(2010), Roman-Duval et al. (2010), Leroy et al. (2011), Bolatto et al.(2011),andSmithetal.(2012).Graypointsshowhighquality solutions from analysis of 22 nearby disk galaxies by (Sandstrom et al. 2012), with typical uncertainties illustrated by the error bars near the bot- tom left corner. Metallicities are from Israel (1997), Bolatto et al. (2008), and Moustakas et al. (2010) and quoted relative to solar in the relevant system (12 + log [O/H] = 8.7forthefirsttwo,12+log[O/H] = 8.5forthelatterwhichusesthemetallicitycalibrationby Pilyugin & Thuan 2005). Note that significant systematic uncertainty is associated with the x-axis. The color bands illustrate our recom- mended ranges in αCO for the Milky Way and ULIRGs. (Right) Colored lines indicate predictions for XCO as a function of metallicity from the references indicated, normalizedOLATTO to XCO,20 OLFIRE=2atsolarmetallicitywherenecessary.ForthesepredictiEROY ANDSTROMons,ET we assumeAL that GMCs −2 B , W , & L (2013), S . (2012) have !ΣGMC" =100M! pc ,whichwetranslatetoameanextinctionthroughthecloudusing Eq. 21. Dust-based determinations find asharpincreaseinXCO with decreasing metallicity below Z ∼ 1/3–1/2 Z!. 1 range of possible power law exponents, XCO Z to 7. XCO IN STARBURSTS AND OTHER LUMINOUS Z3. The strength of this approach is that the∝ observa- GALAXIES tions needed to make such estimates are widely acces- Molecular gas in starbursts exists under conditions sible. The weakness, of course, is that it requires as- very different from those found in most normal galaxies. suming an underlying relationship between H2 and star Observations of starbursts suggest widespread gas vol- formation. Any true dependence of the star formation ume and column densities much higher than those typi- efficiency on metallicity, or any other quantity covariant cal of normal disks (e.g., Jackson et al. 1995; Iono et al. with metallicity, will be recast as additional variations in 2007). Molecular gas in starbursts is also warmer, ex- XCO. citing higher rotational transitions than those found in If XCO does increase rapidly moving to low metallic- less active objects (e.g., Bradford et al. 2003; Ward et al. ity, our knowledge of the distribution function of molec- 2003; Rangwala et al. 2011). In fact, a negative correla- ular column density will present a practical limit to the tion is observed between molecular gas depletion time usefulness of CO to trace H2. At metallicities perhaps (a parameter that characterizes how long a galaxy can as high one half solar, half of the H2 mass will exist maintain its current star formation rate), and excita- outside the CO-emitting surface, and that fraction will tion (e.g., Paglione, Jackson & Ishizuki 1997). Similarly, rapidly increase for decreasing metallicity. Thus, appli- a positive correlation exists between gas density and star cation of a CO-to-H2 conversion factor at very low metal- formation rate (e.g., Gao & Solomon 2004). These ob- licity ultimately involves extrapolating the total mass servations show that there is a fundamental relation be- ! of a cloud from only a small inner part; by Z 0.1 tween the density and temperature of the molecular gas ∼ this may already be analogous to measuring the total and the existence of starburst activity, such that the gas H2 mass of a Milky Way cloud from HCN emission or present in starbursting galaxies or regions has higher den- some other high density tracer.