5 Polarization versus reddening

As mentioned before, the observed sample contains stars for which distance and colour excess have been accurately determined. It is then instructive to compare the spatial distribution of the colour excess and polarization. It must be, however, noted that the polarization values listed in Table 1 are by definition positive quantities, which suffer a positive bias which is not negligible at low polarization levels, while colour excess do not suffer the same effect.

Usually, it is observed that although polarization and extinction (reddening) occurs whenever stellar light is propagated through a medium containing small particles, the correlation between these two quantities is by no means perfect. While the reddening caused by successive clouds is always increasing, the polarization has a more complex behaviour. This fact is illustrated by diagrams where the interstellar polarization is represented against the reddening (e.g. Serkowski et al.1975) which show that the ratio ${P}_{\rm max}/{E}_{B-V}$ rarely exceeds 9.0. Figure 6 (left) gives the polarization versus colour excess diagram for all observed stars. The straight line represents the above mentioned optimum alignment efficiency, which is reached under special conditions only. Figure 6 (right) gives an expanded view for the low reddening part of the polarization versus reddening diagram. As expected, there are several stars showing a ratio P/E_B-V < 9.0, but in general, most of the observed stars are distributed close to the optimum alignment line indicating a rather good correlation between these two quantities.

In Fig. 7 one can compare the distribution of polarization and colour excess as a function of the stellar distance. There are 32 stars up to a distance of 70 pc. This stellar group presents an average polarization of $\overline{{P}_{B}} = 0.03\%$, which is similar to the estimated mean error obtained for them ($\overline{\sigma_{ B}} = 0.02\%$), and an average colour excess of $\overline{ E}(b-y) = 0\hbox{$.\!\!^{\rm m}$}0 13$, which is also comparable to the expected mean error for the colour excess determination (Knude 1978). The obtained average polarization is $\approx 5$ times smaller than the expected maximum polarization value for optimum alignment efficiency. This result enforces the fact that the solar neighbourhood is a low column density volume, at least when concerning the third and fourth galactic quadrants, up to distances of about 70 pc. The colour excess diagram (Fig. 7 - lower) shows clear signs of the existence of interstellar dust at a distance of $\approx 60$ pc. Figures 8-11 give a comparison between the spatial distribution of polarization and colour excess for four distance intervals. In the first interval (d < 100 pc), most of the observed stars are unpolarized and the few ones showing some degree of polarization also show some trace of reddening. On the other hand, in the second distance interval ($100 \le d < 150$ pc) one clearly see the effects of the interstellar dust on the stellar light yielding polarization as well as reddening. The interstellar dust traced here may be related to a large structure defining the interface between the Local and Loop I Bubbles. An extensive study of reddening versus distance conducted by Corradi et al.(1997) suggested the existence of a continuous sheet structure extending from the Chamaeleon region towards the Coalsack ($294\hbox{$^\circ$}\le l \le 308\hbox{$^\circ$}$, $-18\hbox{$^\circ$}\le b \le 5\hbox{$^\circ$}$), at a distance of $150\pm30$ pc from the Sun, which has been associated to this interface.

Acknowledgements

The Danish Board for Astronomical Research is thanked for allocating the observing periods. The Calrsberg Foundation has provided us a grant to cover the expenses of one of the observing missions. We would like to thank the referee, Dr. J.L. Leroy, who provided valuable suggestions to the improvement of this paper. The Brazilian Agencies CNPq and FAPEMIG are acknowledged for partially supporting this research.