Astron. Astrophys. 358, 835-840 (2000)

4. Discussion and conclusions

On the basis of a sample of approximately 50 BAL QSOs, we have investigated possible correlations between BAL QSO radio properties and other spectral characteristics, including polarization and the terminal velocity of the flow.

The main result of our statistical study is the absence of correlations between the radio-to-optical flux ratio, , and all other quantities: polarization and slope of the continuum, balnicity and detachment indices, and the terminal velocity of the flow. The anticorrelation between and claimed by Weymann (1997) and Kuncic (1999) is therefore not confirmed. Furthermore, we do not support the relation between and suggested by the model of Goodrich (1997), even if we restrict our sample to the formally radio-intermediate objects.

The only possible correlations occur within the LIBAL QSO sub-sample, which, once more, seems to behave differently. But the significance is marginal and additional data are necessary. It is nevertheless important to remark that, if real, the relation between and is opposite to the prediction of Goodrich (1997) who suggests that would be higher for objects with large as a result of a stronger attenuation of the UV rest-frame continuum.

It is also interesting to note that is uncorrelated with the continuum power-law index , i.e. with dust extinction (Sprayberry & Foltz 1992). This is in agreement with the results of Hall et al. (1997) who found, from the distribution of optical / near-infrared colours, that the excess of BAL QSOs among radio-intermediate QSOs cannot be attributed to extinction.

Within the unification scheme, different orientation and attenuation of the continuum are probably necessary to explain the different polarization properties of BAL and non-BAL QSOs (Goodrich 1997). Our results indicate that this interpretation cannot simultaneously explain the excess of BAL QSOs among radio-intermediate QSOs. Another interpretation of BAL QSO radio properties is therefore needed.

Alternatively, we may abandon the hypothesis that all observed differences between BAL and non-BAL QSOs are only due to different orientations. The BAL phenomenon may then be seen as an evolutionary mass-loss phase, and properties like polarization could be related to the presence of ejected material. The range of BAL QSO radio properties can also be explained considering models like that of Kuncic (1999) which associates the BAL region with a poorly collimated and weakly radio-emitting jet. The lack of correlation between and does not support this model, although it should be pointed out that it does not necessarily refute it. Indeed, the relation between and may depend on several other parameters, like the covering factor or the orientation of the jet. In this view, the eventual detection of a significant anticorrelation between and within a larger LIBAL QSO sub-sample would fix some of these parameters, and provide useful constraints on the models.

© European Southern Observatory (ESO) 2000

Online publication: June 20, 2000
helpdesk.link@springer.de