STScI, 29 Mar 2012 1

How Stars Die: Infrared Spectroscopy of Dusty Carbon

Stars in the Local GroupG. C. Sloan

A.A. Zijlstra, E. Lagadec, M. Matsuura,

K.E. Kraemer, M.A.T. Groenewegen,

I. McDonald, J.T. van Loon,

J. Bernard-Salas, & P.R. Wood

STScI, 29 Mar 2012 2

Getting from there to here

STScI, 29 Mar 2012 3

The Local Group projectObjective – Understand dust

production of evolved stars as a function of metallicity

Method – Use the Infrared Spectrograph on Spitzer to study carbon stars in nearby dwarf spheroidal galaxies

The great simplification – Treat each of these complex systems as having a uniform metallicity

Results – Sloan et al. (2012, ApJ, submitted)

http://isc.astro.cornell.edu/~sloan/library/

STScI, 29 Mar 2012 4

Samples and metallicitiesGalaxy [Fe/H] ~ 0Large Magellanic Cloud

~ –0.3 D = 50 kpcSmall Magellanic Cloud

~ –0.7 60 kpcFornax dSph

~ –0.3-0.8 150 kpcSculptor dSph

~ –1.0 87 kpcLeo I dSph

~ –1.4 280 kpcCarina dSph

~ –1.7 100 kpc

STScI, 29 Mar 2012 5

Mbol mass age [Fe/H]

Right: Fig. 14 from Revaz et al. (2009), based on evolutionary models

Fornax – Most targets are younger than ~3 Gyr– Metallicities most like SMC and LMC

Sculptor – Both targets are <2 Gyr old – [Fe/H] ~ –1.0

STScI, 29 Mar 2012 6

A carbon starIRAS 05373-0810 (V1187 Ori)

Szczerba et al. (2002)

STScI, 29 Mar 2012 7

Local Group spectra

• These targets are faint!• Need Cornell’s optimal extraction algorithm

(Lebouteiller et al. 2010)• 10,000 extracted spectra publicly available:

http://cassis.astro.cornell.edu

STScI, 29 Mar 2012 8

Manchester Method

Total warm amorphous carbon contentMeasured by the [6.4] – [9.3] colorNeed outflow velocity, gas-to-dust ratio to get mass-loss rateCalibrated with radiative transfer models (Groenewegen et al. 2007)

Gaseous acetylene absorption strength at 7.5 mSiC dust emission strength at 11.3 m

Introduced bySloan et al. (2006) andZijlstra et al. (2006)

Applied to large comparison samples from the Galaxy, LMC, and SMC

STScI, 29 Mar 2012 9

Metallicity diagnostics

Fornax follows the SMC (as expected)

Sculptor and Leo I are (mostly) in the upper leftMAG 29 in Sculptor is off-scale, with EW = 0.8 m and no SiC!(But even that can’t account for the expected free carbon)

In more metal-poor samples:Acetylene bands strengthenSiC dust emission weakens

Leads to a metallicity gradient in the figure

SMC… LMC… Milky Way

STScI, 29 Mar 2012 10

Total mass-loss rates

• [6.4]–[9.3] scales with dust opacity (aka dust content)• Multiply by outflow velocity to get dust-production rate• Multiply by gas-to-dust ratio to get total mass-loss rate

STScI, 29 Mar 2012 11

Carbon-rich dust content

Pulsation periods from the SAAOFornax: Whitelock et al. (2009)Sculptor: Menzies et al. (2011)Leo I: Menzies et al. (2010)

Their work is the key to making these comparisons possible

Dust content increases with pulsation period

Metallicity has little obvious influence

STScI, 29 Mar 2012 12

A closer look

We may be seeing a decrease in dust content at the lowest metallicities

Sculptor and Leo I are below the fitted line, at a 3.6 level (The Fornax data are

consistent with our assumed metallicity)

STScI, 29 Mar 2012 13

C/O and metallicity

After formation of CO molecules

• Assume Ci scales with Z

• Assume C independent of Z

• O = Oi does depend on Z

[O/Fe] = –0.25 [Fe/H]

for –1.5 < [Fe/H] < 0.0 Melendez & Barbuy 2002, Fig. 5

STScI, 29 Mar 2012 14

Expected free carbon

Take (C/O)⊙ = 0.54 and C = 0.56 O⊙

Galaxy [Fe/H] C/O Cfree/C⊙

Milky Way 0.0 1.1 0.19

LMC –0.3 1.4 0.44

SMC –0.7 2.2 0.68

Sculptor –1.0 3.5 0.81

Four times more free carbon in Sculptor than the Milky Way?

It’s not in the dust!And it’s not in the C2H2

STScI, 29 Mar 2012 15

Consequences

• Observation: Little change in amorphous carbon dust content with metallicity (Z)

• But we expect much more free carbon at low Z– Because the 3 sequence and dredge-up should

not depend on Z, and there’s less O to make CO• Conclusion: The dredge-up must be truncated• Consequence: When the free carbon exceeds

some threshold, it triggers a superwind, which strips the envelope, ends life on the AGB, and produces a PN

STScI, 29 Mar 2012 16

Consequences2

The mass-loss history and lifetime on the AGB will determine what a star can produce and inject back into the ISM

STScI, 29 Mar 2012 17

The End