Using IR Emission to Trace H 2 in

Using IR Emission to Trace H 2 in Galaxies CO emission from the SMC (Mizuno+ 01) Adam Leroy NRAO Charlottesville Hubble Fellow Alberto Bolatto Karl Gordon Erik Rosolowsky Snezana Stanimirovic Frank Israel Caroline Bot Andreas Schruba Karin Sandstrom Norikazu Mizuno Evidge Corbelli Elias Brinks Fabian Walter Dust emission from the SMC: SAGE-SMC (PI: K. Gordon), S 3 MC (Bolatto+ 07)

Ways to Get at H 2 (in Other Galaxies)… The H 2 that makes up most of the mass in GMCs is practically invisible. Indirect tracers have to be used, which may have their own environmental biases. H 2 rays Dust CO UV absorption studies: great! But requires background source, probe pencil beam. Rotational line emission (IRS), sensitive only to gas down to T~100 K (most H 2 much colder). Cosmic ray hits nucleus, produces ray… modeled to yield total nuclei column density. Major resolution and sensitivity challenge beyond Milky Way, modeling complex… Probes total gas column modulo dust-to-gas ratio. Absorption tough because of lack of clear background screen. Emission limited by finite resolution of IR telescopes. Next most common molecule after H 2. Standard tracer of H 2 in high-mass galaxies. At low metallicities C and O less abundant and dust shielding weak. CO suppressed? Comes with velocity information - dynamics offer another way to trace mass.

Why We Think CO May Break in Low Mass Galaxies… CO is faint, but when it can be measured (roughly the Local Group) these galaxies have low CO-to-SFR, CO-to-HI and CO-to-starlight ratios. Is this a real effect on the H 2 or a breakdown in our tracer? Dots: measurements from various local galaxies. SFR per unit area CO per B-band Luminosity Dots: relatively massive star-forming galaxies. Local Group Dwarfs (very low CO / SFR) Stellar Luminosity [Magnitudes] CO per unit area

The CO-to-H 2 Conversion Factor… Is the amount of H 2 inferred from various independent tracers consistent? e. g. , CO brightness, dynamics, and dust emission… Ø Ratio of H 2 column density to CO intensity. XCO Ø Certainly a function of scale studied. Ø In this talk: shorthand for “how much H 2 is there? ” Ø Milky Way value ~1. 5 - 3. 0 1020 cm-2 (K km s-1)-1 Ø Recast as , equivalent parameter in mass units.

Using CO Dynamical Masses to Calibrate XCO… Early (often interferometer) studies used dynamical masses and found a moderate increase in the amount of H 2 per CO as metallicity decreases. Arimoto+ ‘ 96 CO-to-H 2 Conversion Factor Wilson ‘ 95 Metallicity GMCs seen in CO in IC 10

Using CO Dynamical Masses to Calibrate XCO… In subsequent studies at higher resolution and sensitivity this effect has diminished. XCO from Virial Mass (units of Milky Way) Points are galaxies, error bars show full range of XCO Linear SMC Milky Way range Metallicity [12+log O/H] See also: Walter et al. (2001, 2003) in NGC 3077, 4214 and Israel et al. (2003) in Magellanic Clouds Bolatto et al. (2008)

Using Infrared Emission Instead of CO… Dust emission offers another independent check on the H 2 distribution. Observed CO o o Better than CO? at least different biases… o In the Galaxy, matches rays and CO well. Predicted from IRAS Dust Emission Dust traces the total gas (HI + H 2) column. Comparison in Profile (pretty good match!) Dame+ 01 (see also Bloemen ‘ 90) H 2 = ( dust × DGR-1)- HI Estimate dust surface density from IR (need at least two bands to make a temperature estimate). Measure the dust-to-gas ratio from the ratio of dust to atomic gas away from the molecular line emission but near enough to calibrate out galactic variations. Subtract the already known distribution of atomic gas.

Using Infrared Emission Instead of CO… Suggests strong XCO variations in dwarfs. Israel (1997) o IRAS towards many galaxies CO-to-H 2 Conversion Factor o temperature correction based on 60/100 o very steep dependence, XCO Z-2. 7 IR-to-HI map of the SMC Metallicity

Using Infrared Emission Instead of CO in the SMC… Higher resolution application to the SMC finds large XCO with H 2 in roughly the same spot as the CO peaks but more extended. 45 pc resolution to match NANTEN CO Dust H 2 peaks match CO, but H 2 more extended Total H 2 mass 3 × 107 Msun much more than implied by CO. H 2 ~ 10% of the HI mass. Roughly in line with SFR and stellar content. Leroy et al. (2007) See also: Rubio+ 04, Bot+ 07; LMC study by Bernard+ 08

Using Infrared Emission Instead of CO in the SMC… Mizuno et al. (2001) - NANTEN CO map of the SMC 45 pc resolution to match NANTEN CO H 2 peaks match CO, but H 2 more extended Total H 2 mass 3 × 107 Msun much more than implied by CO. H 2 ~ 10% of the HI mass. Roughly in line with SFR and stellar content. Leroy et al. (2007) CO

Why the Discrepancy Between Dust and CO Dynamics? The discrepancy between CO, dynamics, and IR may be largely a function of PDR structure, driven by dust shielding… As metallicity and dust-to-gas ratios decrease… … Av~1 mag moves deeper into a given clump. CO disappears when Av<2 mag through a clump… … but H 2 exists to much lower extinctions S. Glover Poster CO kinematics pick up only inner (CO-bright) part of cloud C+ CO H 2 Z Clumps at decreasing (left to right) metallicity… Maloney & Black (1988), Lequeux et al. (1994), Bolatto et al. (1999), Röllig et al. (2006)

Infrared Emission and CO at High Resolution… A similar approach at high (10 pc) resolution in the SMC Wing appears to work… CO Leroy et al. (2009) H 2 from Dust 10 parsec resolution (previous map ~45 pc)

Infrared Emission and CO at High Resolution… CO Intensity On 10 pc scales, the data appear consistent with dust shielding playing a key role in whether or not CO is present… Leroy et al. (2009) Line-of-Sight Extinction (Estimated From IR Emission)

This Experiment… 1 2 3 Measure HI, CO, and IR (24+70+160) over matched lines sight. Try to focus on areas where HI ~ H 2. Estimate the dust surface density (or optical depth) from FIR emission. Need at least two bands (here 70, 160) to make a temperature estimate. Vary XCO and measure scatter about the bet-fit dust-to-gas ratio. Look for XCO that minimizes scatter or maximizes dust-gas correlation.

The Two Nearest Spirals… M 31 HI: WSRT Brinks & Shane ‘ 84 IR: Spitzer Gordon et al. ‘ 05 CO: IRAM 30 m Nieten et al. ‘ 06 M 33 HI: WSRT Deul & van der Hulst ‘ 87 IR: Spitzer K. Gordon see Verley+ ‘ 07, Tabetabei+ ‘ 07 CO: FCRAO + BIMA Rosolowsky+ ‘ 07 Heyer, Corbelli et al. ‘ 04

The Two Nearest Star-Forming Dwarfs… LMC HI: ATCA Kim et al. ‘ 98 IR: Spitzer SAGE Meixner et al. ‘ 06 CO: NANTEN Fukui et al. ‘ 99 SMC HI: ATCA + Parkes Stanimirovic et al. ‘ 99, ‘ 04 IR: Spitzer S 3 MC and SAGE-SMC Bolatto et al. ‘ 07, Gordon et al. ‘ 09, ‘ 10 CO: NANTEN Mizuno et al. ‘ 01

Step 1. Grab regions where HI and H 2 both contribute to the ISM… M 31 M 33 LMC SMC CO Maps (roughly) Region Considered

Step 2: Minimize Scatter in (XCO CO + HI) vs. Dust o Error bars: bootstrapping o Not independent o Bar: 1 range for minimum

Step 3: Compare Results Among Galaxies (XCO) CO-to-H 2 Conversion Factor Each point a different methodology… Colors: how to solve Draine et al. (2007) Using Dust in SINGS Milky Way o Correlation o Rank correlation o Scatter o Median Abs. Dev. Error bars: Bootstrapping Point Shape: dust map o Draine & Li 2007 o Modified BB ( =1. 5) o 70 m = 50% VSG o 160 m intensity Galaxy (Decreasing Metallicity)

Conclusions Ø IR emission from dust can be used as an ISM tracer. Ø Gives a way to estimate otherwise invisible H 2. Ø Significant discrepancies from interferometric CO dynamics. Ø Dust shielding key? CO only at core of clouds? Ø Local Group XCO(IR) ~ constant in M 31, M 33, LMC; SMC high Follow-Up How does dust-to-gas ratio and emissivity of H 2 relate to HI? How does CO as a function of AV vary with location and U? Herschel: Expand Local Group experiment to many galaxies.

Gas-to-Dust Ratio (Draine & Li 2007) Sanity Check on Dust-to-Gas Ratio Draine & Li (2007) models May be a bit biased low? (i. e. , too much dust) Milky Way Trend about right (a bit steep? ) Super-linear agrees with Munoz-Mateos 2009

AV Estimates in the Magellanic Clouds from FIR ~10 parsec resolution

Emission at 160 per unit optical depth The 160 Micron Band Temperature [K]