Are the broad emission lines of quasars affected by gravitational microlensing?

Abstract

Gravitational microlensing by stars in galaxies can affect the continuum flux observed from quasars, but it is usually assumed that the size of the broad emission line region (BLR) is too large to be considerably magnified (or demagnified) by stellar-mass lenses. Although this is true, this does not mean that the radiation from the BLR is unaffected by microlensing. If the BLR has intrinsic structure (such as a large-scale velocity field), parts of it can be stronger magnified than others, leading to an influence on the resulting line profile. Whereas the changes of the line profile by microlensing will be too small to be detectable in a single quasar spectrum, a comparison of the same line in different images of a multiply imaged quasar can yield differences, which (with consideration of a possible difference in the light-travel time between these images) can be attributed to the action of microlensing in one (or more) of the images.

Using two simple models for the geometry and kinematics of the BLR (Keplerian rotation and gravitational infall), we study the resulting line profiles with and without microlensing¹. Whereas our approach is quite general, most of the results presented here apply to image A of the quasar 2237+0305, in which microlensing of the continuum flux has recently been verified (Irwin et al., 1989). The reason for the large interest in QS02237+0305 is mainly due to the fact that the lensing galaxy has a very small redshift, which makes all geometric factors for lensing rather favourable.

Our main results can be summarized as follows: (1) There is an observable influence of microlensing on the broad emission lines of quasars, if the motion of the clouds in the BLR is not random or chaotic, but somehow ordered. (2) BLRs with large-scale rotation are much more sensitive to microlensing than infall models; in particular, even the (mean) redshift of the line can be changed in the case of rotation. (3) For all microlensing models of QS02237+0305 A we have considered, there is an appreciable difference of the lensed and unlensed line profiles. Whereas this difference is larger for the case of Keplerian rotation, even for the infall model it is at a level which should be observable by sensitive spectroscopy of the individual images of this multiply imaged quasar. In the case of a rotating BLR, one might see redshift differences of a single broad emission line which amount to up to 200km/s, and fractional deviations of the lensed from the unlensed profile of typically 10%. In Fig.1, some example profiles are shown with the microlensing parameters of 2237+0305A for Keplerian rotation (top) and gravitational infall (bottom); ζ is a normalized measure of wavelength. In Fig.2, similar profiles are shown corresponding to some other microlensing parameters.

If such differences in line profiles of multiply imaged quasars are observed — and there are some indications for it (Filippenko 1989, Vanderriest 1990) — one may be able to put constraints on the geometry and kinematics of the BLR in those quasars; in particular, random or chaotic motion of the BLR could then be excluded. On the other hand, from a lack of such differences, one might be able to exclude, e.g., a model with large-scale rotation. Furthermore, detecting a significant difference in the broad emission line profiles of a multiply imaged quasar, which cannot be attributed to the time-delay between the images, may provide an alternative detection of microlensing, which does not depend on long-time monitoring of the source. Finally, our results yield a typical value for the variation in redshift (as measured from the broad emission lines) which can be tolerated in multiple quasars before they may be ruled out as lensing candidates.