NGC 1808

   
6 Summary and conclusions

We have presented ISOCAM $\lambda = 5.0$- $16.5~{\rm\mu m}$spectrophotometric imaging of the starburst galaxies M 82, NGC 253, and NGC 1808. The spectrum of all three objects, down to the smallest scales of $\sim$ $100~{\rm pc}$ accessible from our data, exhibit similar characteristics including prominent PAH bands, a featureless continuum most obvious at $\lambda \ga 11~{\rm\mu m}$, and a trough in the 10$~\mu$m region. We securely identified the main emission features detected in the $R \sim 40$ ISOCAM data of M 82 and NGC 253 based on their $R \sim 500$-1000 SWS spectra. The comparison emphasizes the caution that should be exercised when interpreting low resolution data because of potential blends or misidentifications between emission lines from ionized gas and PAH features originating in PDRs, a notable example being [Ne II] 12.81$~\mu$m and PAH 12.7$~\mu$m. Using a simple model combining a template PDR spectrum and a power-law $f_{\nu} \propto (\lambda - 8.5)^{1.5}$, we constructed a representative starburst SED and explored the effects of extinction at MIR wavelengths. Our simulations illustrate the importance of the assumed extinction law (e.g. the widely used Draine 1989 law versus the Galactic Center law of Lutz 1999) and of the intrinsic PAH spectrum (especially the gap between the main 6- $9~{\rm\mu m}$and 11- $13~{\rm\mu m}$ complexes) in shaping the SED of astronomical sources. This complicates the interpretation of PAH ratios as well as extinction measurements relying on the silicate dust absorption at 9.7$~\mu$m (see also Sturm et al. 2000).

As observed previously in a wide range of Galactic and extragalactic sources, the 5- $11~{\rm\mu m}$ spectrum in our galaxies is nearly invariant. The relative PAH intensities exhibit nevertheless measureable and significant variations of $20\%$-$100\%$ which may be attributed to various, possibly interrelated effects including the intensity of the incident radiation field and the PAH size distribution, ionization, and dehydrogenation. In our sample, M 82 probably best illustrates variations of PAH ratios due to an increased fraction of ionized PAHs within the most intense starburst sites and, admittedly speculatively, perhaps also to differences in typical PAH sizes depending on the molecular gas concentrations. The PAH 7.7$~\mu$m L/C ratio in all three galaxies clearly lies in the range observed for pure starburst systems and extends the trend reported previously by Rigopoulou et al. (1999) to lower luminosities.

In contrast, the $\lambda \ga 11~{\rm\mu m}$ region varies most among our sample galaxies and the ISOCAM maps show a comparatively more compact 15$~\mu$m continuum distribution relative to the PAH emission. Strong 15$~\mu$m continuum in M 82 and NGC 253 indicates an important contribution by VSGs contrary to the case of NGC 1808 where the long-wavelength emission is much flatter presumably because of negligible contribution by VSGs. We find, however, that in all three galaxies the 15$~\mu$m continuum and [Ar II] 6.99$~\mu$m line fluxes satisfy a linear relationship. We infer from this that the 15$~\mu$m continuum provides a good indicator of star formation activity in starbursts, complementing the similar results of Roussel et al. (2001b) for galactic disks. In a broader perspective, our galaxies fit well in the trend of increasing ISOCAM 15$~\mu$m/7$~\mu$m ratios with higher levels of star formation activity found among normal disk galaxies and starburst-powered LIRGs/ULIRGs (e.g. Laurent et al. 2000; Roussel et al. 2001b; Dale et al. 2001).

The value of the ISOCAM spectrophotometric imaging of M 82, NGC 253, and NGC 1808 presented in this paper also lies in that it complements existing MIR data with maps of PAH features, fine-structure lines, and continuum components not previously imaged. The poorer angular resolution of ISOCAM compared to that achieved with large ground-based telescopes is compensated by the larger field of view revealing more of the large scale emission. In that respect, we stress the small size of the MIR source relative to the optical extent of all three galaxies: the ISOCAM maps cover the central $\approx$ $1.5~{\rm kpc}$for M 82, 0.6 kpc for NGC 253, and 5 kpc for NGC 1808 while the optical diameters are about 10 kpc for M 82 and 20 kpc for the other two. By measuring the flux density within the ISOCAM LW10 filter bandpass, equivalent to the IRAS 12$~\mu$m band, we recover all of the IRAS 12$~\mu$m emission in the entire field of view for M 82 and NGC 1808, and about 70% for NGC 253. This suggests that the total MIR emission is strongly dominated by the starburst sites in the nuclear regions while the more quiescent star formation taking place in the disk at larger radii does not contribute much. It will be interesting to see whether the MIR cameras on board SIRTF, to be launched in 2003, will confirm this result or discover faint diffuse emission, especially at shorter (3-5  ${\rm\mu m}$) wavelengths.

Acknowledgements

We are grateful to E. Sturm for making the SWS spectrum of NGC 253 available to us in electronic form. We warmly thank H. Roussel, A. Vogler, S. Madden, and especially D. Tran for many interesting discussions on various aspects of this work. We also wish to thank the referee, Dr. N. Bergvall, for his valuable comments that improved the quality of the paper. VC would like to acknowledge the partial support of JPL contract 960803.

NGC 1808