3.1. Comparisons with Dust Grain Models
One way to synthesize our knowledge of dust is to fit different observations of dust to a dust grain model. The main observational constraints for dust grain models are usually the diffuse ISM dust extinction curve and dust abundances (Weingartner & Draine 2001; Clayton et al. 2003). In addition, some dust grain models also fit the dust polarization (Kim & Martin 1994; Li & Greenberg 1997) and infrared emission spectrum (Zubko et al. 2003; Li & Draine 2001). Some dust grain models also have been fit to the denser dust extinction curves (Kim, Martin, & Hendry 1994; Weingartner & Draine 2001). No dust grain models have used the measured dust scattering properties as fitting constraints, but most do use the scattering properties as a consistency check.
In Figs. 1 - 4, predictions for a and g from recent dust grain models are plotted. The Zubko et al. (2003) model plotted is described as BARE-NC-B. All but one of the five model predictions plotted are for diffuse ISM dust (RV = 3.1) and are then only directly comparable to the DGL dust scattering results. As can be seen from Fig. 3, the dust grain models are in fairly good agreement with the results for g, but most of the models underpredict a for most wavelengths. All the models have a small dip in a at the 2175 Å bump, but the dip is not as large as indicated by the DGL measurements. The models all predict lower far-ultraviolet albedos than have been measured in the DGL.
The one dust grain model for denser dust (Weingartner & Draine; RV = 5.5) is directly comparable to the dust scattering measurements in reflection nebulae (Fig. 1) and dark clouds (Fig. 2) as these objects have similar RV values. Similar to the DGL comparison, the g values predicted by the model are in reasonable agreement for reflection nebulae and dark clouds, except possible in the optical for dark clouds. The model predictions for a agree with the measurements for dark clouds which mainly probe the optical and near-infrared. The a predictions for reflection nebulae are generally too high in the optical and too low in the far-ultraviolet. The albedo dip at 2175 Å is much smaller in the model than the well measured values indicate. Basically, the recovery of the albedo at around 1500 Å is not seen in any of the dust grain models.