Astron. Astrophys. 352, 371-382 (1999)

1. Introduction

The problem of internal dust extinction in galaxies is difficult whereas its estimate is crucial for understanding the evolution of the Star Formation Rate from high redshift to now. At high z the emission observed in the visible corresponds to the UV rest frame where the effects of the dust extinction can be dramatic. For example the shape of the Madau plot (e.g. Madau et al. 1998) depends a lot on the extinction adopted as a function of the redshift. Since the work of Calzetti, Kinney and collaborators from IUE data (e.g. Kinney et al. 1993, Calzetti et al. 1994) the slope of the UV continuum in the range 1200-2600 Å () has been identified as a powerful indicator of the dust extinction. The reason is that only short lived stars contribute substantially to the emission in this wavelength range and the intrinsic shape of the spectrum is only sensitive to the recent star formation history. For example reaches a steady value as soon as the star formation rate has been constant for some years (Calzetti et al. 1994). An instantaneous starburst represents the most extreme case of a steep spectrum with equal to -2.7 (Meurer et al. 1995). Once an intrinsic value for is adopted any deviation from this value is interpreted in terms of dust extinction which flattens the intrinsic slope. At high redshift is observable in the visible wavelength range and this has conferred a large interest to this approach.

Nevertheless, at least two difficulties arise when using this method: on the one hand the choice of an intrinsic UV slope (i.e. of a star formation history) can modify substantially the amount of extinction and has led to some discrepancies in the estimate of the extinction which can reach 1-2 mag in UV (Pettini et al. 1998, Meurer et al. 1997, Steidel et al. 1999, Calzetti 1997); on the other hand if the deviation of from its intrinsic value is indubitably a dust extinction tracer, deriving a quantitative value of the extinction from this deviation is difficult due to the various unknown factors like geometry or dust properties intervening in the estimate of the extinction (e.g. Calzetti et al. 1994). The quantification of the extinction is easier on nearby galaxies which can be obviously used as templates. The empirical approach of Calzetti and collaborators (Calzetti et al. 1994, Kinney et al. 1994, Calzetti 1997) had the advantage of providing a global attenuation curve for starburst galaxies accounting for geometrical effects in a statistical way.

Another powerful approach lies in making global energetic considerations. Indeed, for nearby templates, it is possible to perform a total energetic budget since the dust emission of these galaxies is almost always known from IRAS observations. Such considerations have led to quantitative estimates of the extinction (Buat & Xu 1996, Meurer et al. 1999). Recently, Meurer et al. (1999) have related the FIR to UV flux ratio and the UV slope to the extinction at 1600 Å. Their unreddened UV spectrum has a slope of -2.23 intermediate between a constant star formation rate and an instantaneous burst.

Uncertainties about the UV extinction are already present at low redshifts. Meurer et al. (1999) find an extinction around 1.8 mag at 1600 Å for the starburst templates observed by IUE whereas we find 1.3 mag at 2000 Å for a sample of nearby starburst galaxies (Buat & Burgarella 1998) and around 0.8 mag for more quiescent disk galaxies (Buat & Xu 1996). The difference is likely to be due at least in part to the properties of the individual galaxies used for these studies but also to different assumptions about the dust absorption as it will be discussed below.

We need to know how to correct individual galaxies for extinction but also how the properties of these individual cases can be extrapolated to the entire population of galaxies. This problem is especially important at high z since as we go farther only the brightest objects become visible. With the availability of the luminosity functions at various wavelengths we have now the possibility to test if the results deduced from the properties of individual galaxies are representative of the mean characteristics deduced from the local luminosity functions.

The basic idea of this paper is to compare the FIR (60 and 100 µm) and UV(0.2 µm) of a sample of nearby galaxies for which selection biases are well known. The sample will be FIR selected and the aim is to study how much these individual galaxies thus selected trace the mean properties of the local universe. After a presentation of our IRAS/FOCA sample (Sect. 2), we discuss the FIR and UV properties of the individual galaxies in terms of extinction and selection biases in Sect. 3. The Sect. 4 is a comparison with the luminosity functions at both wavelengths. In Sect. 5 we derive quantitative star formation rates from the FIR and UV emissions. Endly, in Sect. 6 we compare the FIR and UV properties of Ultra Luminous Infrared Galaxies both at low and high redshift and of the ISOCAM detections of intermediate redshift galaxies in a CFRS field with those of our IRAS/FOCA sample of nearby galaxies.

© European Southern Observatory (ESO) 1999

Online publication: December 2, 1999
helpdesk.link@springer.de