5 Conclusions

ISOPHOT observations were used together with IRAS data to map the far-infrared emission in the starburst galaxy NGC 253. The resulting images were then analyzed with modified blackbody fits as well as radiative transfer modeling. The total FIR luminosity consists of two parts: a warm component ( $T \sim 47{-}60$ K), dominated by the central regions, and a cold component ($T \sim 20$ K) produced in the disk and the halo. Dust temperatures and masses derived from single-temperature fits of the SED integrated over the whole galaxy are in agreement with those obtained by solving the radiative transfer equations. The results obtained in the different regions may be summarized as follows:

Nucleus: the nucleus produces half of the total FIR luminosity of NGC 253. Our fit for the core correctly predicts the observed 1.3 mm continuum emission given by Krügel et al. (1990). The low inferred star formation rates suggest that the nuclear starburst in NGC 253 is in a late phase, having passed a rapid decrease of the star formation rate.

Disk: the resulting dust temperatures are typical for the general diffuse ISM in spiral galaxies. The radiative transfer model indicates that the dust scale length is $\sim$40% larger than that of stars; the same trend was found by Alton et al. (1998b). The dust is warmer (T > 20 K) in the central regions of the galaxy, the outer regions are dominated by colder dust ($T \le 15$ K).

Halo: we find evidence for extended emission to projected distances of $\sim$10 kpc from the nucleus along the minor axis both in the IRAS and ISOPHOT maps; the emission seen in the IRAS maps could be due to instrumental effects, but since the same emission pattern is also seen in the ISOPHOT map as well as in ROSAT images (e.g. Pietsch et al. 2000; Forbes et al. 2000) we conclude that this is not the case. The SED fitting of this extended component is very uncertain, since IRAS fluxes may still be affected by instrumental effects. Keeping in mind this uncertainty, the emission is probably dominated by cold dust ( $T \sim 15{-}20$ K); the derived dust masses are comparable to those found for the outflows in NGC 1808 and M 82. The apparently constant mass in the outflows observed in nearby galaxies is probably due to limitations in the instrumental resolution and sensitivity. It was not possible to reproduce the halo component by a single 3-D exponential dust disk. Even if in our opinion this is most likely due to a non exponential distribution of dust in the halo, some extra heating may be given by star formation, which according to Comeròn et al. (2001) may occur in the halo. Both phenomena may be related to the interaction of the interstellar medium with the superwind driven by the nuclear starburst.

Acknowledgements

The ISOPHOT development and the postoperation phase performed in the ISOPHOT Data Centre at the Max-Planck-Institut für Astronomie Heidelberg are supported by Deutsches Zentrum für Luft- und Raumfahrt (DLR), Bonn. The authors are responsible for the content of this paper. PIA has been jointly developed by the ESA Astrophysics Division and the ISOPHOT consortium. We are grateful to the referee for the useful comments which improved this paper. We are indebted to Manfred Stickel of the Heidelberg ISOPHOT Data Center for having provided his procedures. We thank IPAC for the IRAS HIRES maps. This research has made use of the NASA-IPAC extragalactic database (NED) which is operated by the Jet Propulsion Laboratory, Caltech, under contract with the NASA.