Star Formation: Near and Far Neal J. Evans II with Rob Kennicutt

Far: Whole Galaxy Relations Solid circles are disk- averaged normal spirals Open circles are central regions of normal disks Squares are circumnuclear starbursts Slope is 1.4±0.15 Kennicutt 1998, ARAA 36, 189 StarburstsSpirals

black: normal galaxies red: starbursts green: circumnuclear starbursts blue open: Low metals (<~1/3 solar), mostly dwarfs Blue line: slope of 1.4, not a fit RCK, in preparation

SFR/Mass Increases with SFR  SFR/Mass of molecular gas increases with SFR  Factor of ~ 100  “Efficiency” increasing  But what does this really mean?  Solomon & Vanden Bout (2005 ARAA) S tar formation efficiency Star formation Rate

The Dense Gas SF Relation  L FIR correlates better with L(HCN)  Smaller scatter  Higher rate  SFR rate linearly proportional to amount of dense gas  “Efficiency” for dense gas stays the same  Gao & Solomon (2004) ApJ 606, 271 Amount of dense molecular gas Star formation rate

Whole Galaxy Prescriptions  Kennicutt (1998)   SFR (M sun yr –1 kpc –2 ) = 2.5x10 –4   gas (M sun pc –2 )  Gao and Solomon (2004)  SFR (M sun /yr) ~ 1.8 x 10 –8 M(dense) (M sun )   SFR (M sun yr –1 kpc –2 ) = 1.8x10 –2   dense (M sun pc –2 )

What Does  SFR Mean?   SFR is grand average over:  Whole galaxy, with huge variations in  SFR,  gas, metallicity, …  Time  ~5 Myr for H   ~ Myr for UV,  Myr for FIR (short for starbursts)

What Does  gas Mean?  “  gas ” is not the mean surface density of any structure.  At best, the filling factor x mean cloud emission times X(CO)  Higher “  gas ” really means more clouds in beam

CO: Limited Dynamic Range Heiderman et al CO can be off by large factors in some regions. It clearly fails for A V > 10 mag. Need A V >0.4 mag for CO, but issues below 3 mag (Pineda et al. 2010)

Not so Bad on Average  12 CO underestimates A V at  gas > 200 M  pc –2 by 30%  Constant value of 13 CO vs  gas, underestimating  gas by factors of 4-5  Correcting for 12 CO, would flatten the slope of the Kennicutt-Schmidt relation (but does not explain big offset)

Intermediate: Resolved Studies  Radial cuts or averages  Martin and Kennicutt (2001): threshold  Schruba et al. (2011): SF continues even when HI > H 2  Pixel by pixel: e.g.,  Kennicutt et al. (2007)  Bigiel et al. (2008)  Blanc et al. (2009)

Sub-kpc scales Bigiel et al Study of 18 nearby galaxies with sub-kpc resolution in HI, CO. SFR from UV+24 micron Threshold around 10 M sun pc –2 in total gas: transition from HI to H 2

CO, SF continue into HI region Schruba et al SFR ~ I(CO) even in HI dominated outer parts

Star Formation Prescriptions for sub-kpc scales  Kennicutt et al. (2007) M51   SFR (M sun yr –1 kpc –2 ) = 1.7x10 –4   37 mol (M sun pc –2 )  Bigiel et al. (2008)   SFR (M sun yr –1 kpc –2 ) = 7.9x10 –3   0 mol (10 M sun pc –2 )   SFR (M sun yr –1 kpc –2 ) = 7.9x10 –4   0 mol (M sun pc –2 )  Blanc et al. (2009) M51   SFR (M sun yr –1 kpc –2 ) = 5.1x10 –2  0.82 mol (M sun pc –2 )  Includes 0.43 dex scatter in  SFR and includes limits  Issues of tracer, diffuse emission, fitting method

Star Formation Prescriptions Theory  Schmidt (1959)  SFR ~  n, n = 1 or 2 (or  n, 2009)  Krumholz et al. (2009)   SFR = f(  gas, f(H 2 ), Z, clumping)  Nearly linear with  mol below ~ 100 M sun pc –2  Steepens above 100 M sun pc –2  Other dynamical relations

The Predictions

17 Very Near: Clouds in Solar Neighborhood Spitzer Programs c2d + Gould Belt: 20 nearby molecular clouds (blue circles) Cluster Project: 35 young stellar clusters (red circles) 90% of known stellar groups and clusters within 1 kpc (complete to ~ 0.1 M Sun )

Whole Clouds (2-16 pc) Heiderman et al Almost all clouds within 300 pc Total SFR from YSO counting /area Total mass/area

Clouds within 1 kpc Adds Orion, Mon R2, S140, Cep OB3, all forming more massive stars, and North America nebula, less active not complete to 1 kpc, but representative

It’s Worse than that… Gray is extinction, red dots are YSOs, contours of volume density (blue is 1.0 M sun pc –3 ; yellow is 25 M sun pc –3 )

Really Near: Within Clouds Heiderman et al. 2010

Less Near: Add Clouds to 1 kpc Gutermuth et al. subm. N = 2.67 N = 1.87

Cep OB3 Gutermuth et al. subm.

Still Less Near: Dense Clumps L(HCN J = 1-0) L(IR) Wu et al. (2005) Survey of dense clumps across MW. (n ~ 10 5 to 10 6 cm –3 ) Birthsites of large clusters. Follow linear relation very similar to dense gas relation for starbursts, as long as L FIR > L sun.

Dense Clumps on  gas -  SFR Using L FIR to get SFR, likely underestimates. Includes fit from Wu et al.

Combine with Nearby Clouds Fit with broken powerlaw with slopes of 4.6 below and 1.1 above a turnover  gas = M sun pc –2. (see Lada et al. 2010) Gutermuth et al. favor continued rise with  SFR ~  gas 2 throughout. All agree: well above all exgal relations except for dense gas relation.

Lessons from Nearby Clouds   SFR >10 times prediction of relations for galaxies  SFR determined on sub-pc scales << exgal resolution  On scales where SF actually happens…  Dependence on S mol may be very strong, at least up to S mol ~ 100 M sun pc –2

Speculation  The underlying SF law is linear in  gas above a noisy threshold ~ 100 M sun pc –2  10 times exgal relations around threshold.  Fraction of gas above threshold (f dense ) increases with as  0.5 for >100 M sun pc –2  When ~ 100  th, f dense ~1  KS prescription and Dense gas prescription agree  What about linear relations in resolved studies of non- starbursts?  f dense ~ constant below ~ 100 M sun pc –2 ?

Issues for Resolved Studies  SFR have be restricted to local SF  Remove diffuse emission  Use tracer with short timescale  Clouds are not resolved, much less clumps  “  gas ” is still not that of any structure  Small number statistics cause larger spread  Massive stars can destroy clouds  SF tracers and gas may even anti-correlate

30 Observe the Solar Neighborhood from Outside Size and location of beam/pixel causes huge variations All centered on Sun 100 pc: No SF, no CO 300 pc: SF, CO, but no H , little 24  m 500 pc: SF, H , CO

What would we see? 300 pc, count YSOs, 500 pc, count YSOs 300 pc, using H , remove diffuse emission 500 pc, using H , remove diffuse emission, assume standard L(H  ) to SFR Bigiel et al. 2011

The Larger Context of MW  Surveys in mm continuum finding 1000’s of dense clumps  Bolocam Galactic Plane Survey (>8000 sources)   ATLASGAL survey from APEX  Future SCUBA2 survey  Herschel Galactic Plane Survey (HIGAL)  Infrared Dark Clouds (IRDC)  MSX, GLIMPSE, MIPSGAL  New models of Galaxy, VLBA distances, …  Provide link to extragalactic star formation

The Improved Milky Way Model Green and blue dots show VLBA measurements of distance, which align star-forming regions along spiral arms much better than previous distances.

Summary  Star formation highly concentrated to dense regions  Steep increase in  SFR to at least  gas > 120 M sun pc –2  x more SF than predicted by any prescriptions  SFR ~ Mass of gas above a threshold density  Non-linear nature of KS relation:  A consequence of f dense ~ 0.5 ?  Resolved studies of galaxies must watch for systematic issues

Backup Slides

A Popular Explanation for Non-linear Relation  Free-fall time depends on volume density  t ff ~ r –0.5  Common theoretical approach  Krumholz and Thompson  Narayanan et al.  SFR ~ Mass/t ff  d r * /dt ~ r/r –0.5 ~ r 1.5  Local version of Kennicutt relation

Any evidence for this? Mean density from virial mass and radius of well-studied sample of dense clumps. ~ M/r 3 (Wu et al. 2010) ~ SFR

Nor in YSO Counting Yellow stars are from Class I and Flat SED SFRs in c2d+GB Clouds.

Milky Way Estimates  Volume filling factor of molecular gas (as traced by CO) is about (Heyer, prelim estimate)  Volume filling factor of clumps (density of few x 10 3 cm – 3 ) < 10 –5 (M. K. Dunham, prelim estimate)