Next Contents Previous


2.1 Variations Abound ...

Figure 1 shows several estimates of the shape of a ``mean'' Galactic extinction curve from the far-IR through the UV region. The curves are presented in a commonly used normalization scheme, E (lambda - V) / E (B - V), and plotted against inverse wavelength. The significance of the different curves will be discussed later in this paper. For the present, these curves serve to illustrate the overall characteristics of interstellar dust extinction. Typically, the extinction rises through the IR with a power law-like dependence (see Appendix), rolls over slightly in the optical region (``knee''), shows a prominent feature at 2175 Å in the near-UV (``bump''), and has a sometimes steep rise in the far-UV (``fuv rise''). Figure 1 does not attempt to show the relatively narrow optical/IR features which may be associated with interstellar dust, such as the diffuse interstellar bands in the optical or the solid state absorption (or emission) features in the IR (e.g., the 9.8 µm silicate feature).

Figure 1

Figure 1. Normalized interstellar extinction curves from the far-IR through the UV. Several general features of the curves are noted. The solid and dotted curves are estimates for the case R ident A (V) / E (B - V) = 3.1 derived in the Appendix of this paper and by Cardelli et al. 1989, respectively. The dashed curve shows the average Galactic UV extinction curve from Seaton 1979.

IR and optical extinction traditionally have been studied using ground based broad- or intermediate- band photometry and sightline-to-sightline variations have been recognized for a long time (e.g., Johnson 1965). These are often characterized using the parameter R ident A (V) / E (B - V), i.e., the ratio of total to selective extinction at V. The value of R ranges between about 2.2 and 5.8 for sightlines along which UV extinction has also been measured, with a mean value of ~ 3.1 for the diffuse interstellar medium. It is believed that the shape of the IR extinction law longward of ~ 8000 Å (the ``power law'' region) may be invariant and that the observed range in R is due to spatial variations in the steepness of the extinction in the optical region around the ``knee'' (e.g., Cardelli et al. 1989; Martin & Whittet 1990).

Measurements of UV interstellar extinction have relied on satellite or rocket spectrophotometric data, and date back to the original discovery of the 2175 Å bump nearly 30 years ago (Stecher 1969; Bless & Savage 1970). These studies immediately revealed that sightline-to-sightline differences exist within the Galaxy in the detailed properties of UV extinction curves. Important results on the degree of spatial variability of UV extinction were subsequently derived from the large photometric databases accumulated by the TD-1 and ANS satellites (e.g., Koornneef 1978; Kester 1981; Meyer & Savage 1981).

The most graphic illustrations of spatial variations in UV extinction were provided by the IUE satellite. With its complete spectral coverage between 1150 and 3200 Å and relatively high spectral resolution (~ 6 Å in the low-resolution mode) IUE was well-suited for the study of UV extinction and many studies have utilized IUE data to determine the wavelength dependence of extinction along one or more Milky Way lines-of-sight (e.g., Bohlin & Savage 1980; Witt et al. 1981; Seab et al. 1981; Morgan et al. 1982; Hecht et al. 1982; Massa et al. 1983; Witt et al. 1984; Massa & Savage 1984; Franco et al. 1985; Aiello et al. 1988; and many others). Figure 2 shows analytical representations for ~ 80 UV extinction curves derived from IUE data (solid curves), illustrating the wide range of properties observed for Galactic UV extinction. These curves are taken from the catalog of Fitzpatrick & Massa (1990), with a few additions, including the lines-of-sight toward HD 210121 (Welty & Fowler 1992) and HD 62542 (Cardelli & Savage 1988). The basis for the analytical representation will be discussed further below.

Figure 2

Figure 2. Examples of 80 Galactic UV extinction curves derived from IUE satellite observations. Analytical fits to the curves are shown, based on the work of Fitzpatrick & Massa 1990. The curves are taken from the Fitzpatrick & Massa catalog, with the addition of the lines-of-sight toward HD 210121 from Welty & Fowler 1992 and HD 62542 from Cardelli & Savage 1988. This figure demonstrates the enormous range of properties exhibited by UV extinction in the Milky Way. The dotted line, labeled ``sigma,'' shows the standard deviation of the sample scaled to the value sigma(1500 Å) = 0.74, as derived from ANS satellite data (see Section 3.1).

The overall impression that has come from the last 30 years of extinction studies is that there is a bewildering variety of IR-through-UV extinction curves and that what has often been called ``peculiar'' extinction is actually quite common. This bodes poorly for the prospects of precisely dereddening astronomical energy distributions. In particular, the range of extinction properties seen in the UV (Figure 2), where the effects of extinction are strongest, may represent a bonanza of information for those who study the properties of interstellar dust grains, but is a potential disaster for those whose main interest in UV extinction is the removal of its effects.

Next Contents Previous