Toward a Quantitative Interpretation of Mid-IR Spectra of Galaxies B. Nikolic (NRAO), P. Alexander (Cambridge), M. Clemens (Padova), D. Ford (Cambridge), G. Cotter (Oxford), M.S. Longair (Cambridge) Figure A: The model correlation between 8-micron luminosity (convolved with the Spitzer filter response) and H-alpha luminosity (broken line styles) for dust shell radii of 100pc, 2kpc and 5kpc (dashed, dotted and dash-dotted respectively). Also shown is best-fitting correlation by Wu et al (2005, solid line and inset). Figure B: As Figure A but showing the expected and observed correlations between 24-micron luminosity and H-alpha luminosity. The Method Starburst99 + Comprehensive dust model Simple radiative transfer + Allows investigation of the effects of star-burst age, metallicity, IMF etc. Also, can directly derive other parameters such as ionising photon flux, S/N rate. Takes fully into account stochastic heating for a full size distribution of carbonaceous and silicate grains. Appropriate for mid-IR models since only UV/optical photons are important in exciting the emission. The model spectra were produced for a star-formation rate of 7 solar masses per year (derived from Brackett line measurements, radio measurements), and with dust shell radius of 400pc (derived from mid-IR imaging above). Galactic dust model Dust particles 3.5-6 A destroyed 7.9 micron MICHELLE image K-band + ISO Case study: NGC 520 ISOCAM/CVS spectra (solid line) Assumed star formation rate 120 solar masses per year (Bracket lines, FIR luminosity), dust shell radius 150pc (mid-IR imaging), column density 1.210 22 cm -2 (Brackett lines). Galactic dust model Dust particles 3.5-35 A destroyed 11 micron MICHELLE image HST/ Thompson 2005 Case study: Arp 220 Spitzer/IRS spectra (solid line) Summary ➔ We have constructed a framework for modelling mid-IR emission of galaxies which makes it possible to directly relate observed spectra or photometry to the star-formation rate, dust properties and/or geometry. ➔ The models naturally reproduce the observed correlation between mid-IR luminosity and star-formation rate. They indicate that dust responsible for the bulk of mid-IR emission is not directly associated with the young stars. ➔ More detailed studies of a moderate (NGC 520) and very intense (Arp 220) starburst are presented. Although there is a rough agreement between models and observations, it is clear properties of the dust must be changing. ➔ The effect of destruction of the smallest grains is investigated and shown to be consistent with the observations. A reasonable fit to the observed spectrum of Arp 220 requires destruction of a large fraction of the small grain population. Acknowledgements & References ➔ We have made use of observations obtained from the ISO and Spitzer archives. Additionally, the results here are based on observations with UKIRT/MICHELLE. ➔ We would also like to thank Dr. Kotilainen for providing us with the K-band image of NGC 520. ➔ The dust model is based on the work of Draine, Li and Weingartner (published in a series of three articles in ApJ in 2001). The Starburst99 stellar population synthesis models are described by Leitherer et al, 1999, ApJS. ➔ The mid-IR/SFR correlations are from Wu et al, 2005, ApJ. Figure C: Strengths of the 3.3 (dotted line), 7.7 (solid line) and 11.3 (dashed line) micron emission features as functions of shell gas column density. The exciting spectrum is from a continuously star forming starburst. Figure D: Evolution of mid-IR emission with starburst age, shown at ages of 5, 100, 475 and 1000 million years. Top is absolute mid-IR luminosity, while in the bottom the traces have been normalised by the stellar luminosity at that age.

Toward a Quantitative Interpretation of MidIR Spectra of

Download PDF Report

Upload
others
View
10
Download
0

Embed Size (px)

Citation preview

Page 1: Toward a Quantitative Interpretation of MidIR Spectra of

Toward a Quantitative Interpretation of MidIR Spectra of Galaxies

B. Nikolic (NRAO), P. Alexander (Cambridge), M. Clemens (Padova), D. Ford (Cambridge), G. Cotter (Oxford), M.S. Longair (Cambridge)

Figure A: The model correlation between 8micron luminosity (convolved with the Spitzer filter response) and Halpha luminosity (broken line styles) for dust shell radii of 100pc, 2kpc and 5kpc (dashed, dotted and dashdotted respectively). Also shown is bestfitting correlation by Wu et al (2005, solid line and inset).

Figure B: As Figure A but showing the expected and observed correlations between 24micron luminosity and Halpha luminosity.

The Method

Starburst99 + Comprehensive dust model

Simple radiative transfer+

Allows investigation of the effects of starburst age, metallicity, IMF etc. Also, can directly derive other parameters such as ionising photon flux, S/N rate.

Takes fully into account stochastic heating for a full size distribution of carbonaceous and silicate grains.

Appropriate for midIR models since only UV/optical photons are important in exciting the emission.

The model spectra were produced for a starformation rate of 7 solar masses per year (derived from Brackett line measurements, radio measurements), and with dust shell radius of 400pc (derived from midIR imaging above).

Galactic dust model

Dust particles 3.56 A destroyed

7.9 micron MICHELLE image

Kband + ISO

Case study: NGC 520ISOCAM/CVS spectra (solid line)

Assumed star formation rate 120 solar masses per year (Bracket lines, FIR luminosity), dust shell radius 150pc (midIR imaging), column density 1.21022 cm2

(Brackett lines).

Galactic dust model

Dust particles 3.535 A destroyed

11 micron MICHELLE image

HST/ Thompson 2005Case study: Arp 220Spitzer/IRS spectra (solid line)

Summary

➔ We have constructed a framework for modelling midIR emission of galaxies which makes it possible to directly relate observed spectra or photometry to the starformation rate, dust properties and/or geometry.

➔ The models naturally reproduce the observed correlation between midIR luminosity and starformation rate. They indicate that dust responsible for the bulk of midIR emission is not directly associated with the young stars.

➔ More detailed studies of a moderate (NGC 520) and very intense (Arp 220) starburst are presented. Although there is a rough agreement between models and observations, it is clear properties of the dust must be changing.

➔ The effect of destruction of the smallest grains is investigated and shown to be consistent with the observations. A reasonable fit to the observed spectrum of Arp 220 requires destruction of a large fraction of the small grain population.

Acknowledgements & References➔ We have made use of observations obtained from the ISO and Spitzer

archives. Additionally, the results here are based on observations with UKIRT/MICHELLE.

➔ We would also like to thank Dr. Kotilainen for providing us with the Kband image of NGC 520.

➔ The dust model is based on the work of Draine, Li and Weingartner (published in a series of three articles in ApJ in 2001). The Starburst99 stellar population synthesis models are described by Leitherer et al, 1999, ApJS.

➔ The midIR/SFR correlations are from Wu et al, 2005, ApJ.

Figure C: Strengths of the 3.3 (dotted line), 7.7 (solid line) and 11.3 (dashed line) micron emission features as functions of shell gas column density. The exciting spectrum is from a continuously star forming starburst.

Figure D: Evolution of midIR emission with starburst age, shown at ages of 5, 100, 475 and 1000 million years. Top is absolute midIR luminosity, while in the bottom the traces have been normalised by the stellar luminosity at that age.