5 Polarization vs. phase angle curves in NIR

The simulations at $1.24~\mu {\rm m}$, $1.65~\mu {\rm m}$ and $2.2~\mu {\rm m}$ are shown in Fig. 6 along with the observations by Hasegawa et al. (1999) and Jones & Gehrz (2000). Except near $30 ^\circ$ phase, the fitted curves do not match very well with these high quality observations. The possible reasons could be any of the following: the narrow size range population of grains in Table 3 which comprise of $61 \%$ in the coma and $12\%$ in the shell sampled by the aperture may not be isolated but may be sub units of highly porous larger grains with porosity $p>90 \%$ as suggested by Kolokolova et al. (2001), Levasseur-Regourd et al. (1997) and Xing & Hanner (1997). Such a grain model is similar to the ballistic cluster aggregation (BCCA) investigated by Mukai et al. (1992). At these porosities, the individual units will interact independently with the incoming visible light and will act as isolated scatterers (Xing & Hanner 1997). However, at infrared wavelengths, the separation between the sub units becomes significant and the porous grain should be treated as an aggregate with non Rayleigh inclusions of these sub units. For such a study the present method of EMT and Mie theory is inadequate and the more elegant technique of Discrete Dipole Approximation (DDA) using the code DDSCAT by Draine & Flatau (1994) will be appropriate. Wolf et al. (1998) compared the results using different EMT rules along with Mie scattering with that using the technique of Discrete Dipole Approximation and showed that significant differences are noticed for grains with non Rayleigh inclusions. In particular, these authors point out that for such grains, computations using DDA reduce forward scattering and the polarization phase function becomes distinctively positive compared to Rayleigh inclusions and the EMT solutions. Large negative polarization at low phase angles in Fig. 6 may significantly reduce if DDA is used. Further, the present method treats the small grain population as individual grains. In the $1.25~\mu {\rm m} {-} 2.2~\mu {\rm m}$ region, these will scatter poorly and the signature of the larger grain population 'b' dictates the resultant polarization at these wavelengths. The present investigation restricted the fit to polarization in the optical region only. Extension to polarization in the JHK bands are important because it will help in investigating the crystal field band of Fe²⁺ in the $1.1~\mu {\rm m}$ region of Fe rich olivines. (Dorschner et al. 1995). This band is week in Mg poor pyroxene. The effect of this band which is spread over the JHK bands will be to increase the opacity of the grains and hence the polarization. A detailed fit from UV to JHK using DDA is planned as the next phase of investigations.