Astron. Astrophys. 358, 793-811 (2000)

8. Summary and conclusions

We have shown unambiguous evidence of external variability in the CLASS gravitational lens B1600+434. The difference between the 8.5-GHz VLA light curves of the two lensed images shows external variability at the 14.6- confidence level. The modulation indices of the short-term variability are 2.8% for image A and 1.6% for image B. The difference light curve has an rms scatter of 2.8%, indicating that the short-term variability in both light curves is mostly of external origin (Sect. 2).

We have investigated two plausible sources of this external variability: (i) scattering by the ionized component of the Galactic interstellar medium (ISM) and (ii) microlensing by massive compact objects in the bulge/disk and halo of the lens galaxy.

Based on the `standard' theory of scintillation (e.g. Narayan 1992; Rickett et al. 1995) there should be a considerable increase in the modulation-index with wavelength (Sects. 3 and 7). From simultaneous WSRT 1.4 and 5-GHz observations we find, however, that =1.2% and =3.7% (Table 1), which is a considerable decrease. Scintillation theory predicts =9.0% for =3.7% (Sect. 7). If the 1.4 and 5-GHz short-term variability is intrinsic, it is hard to reconcile with the fact that in 1998 the VLA 8.5-GHz light curves were dominated by external variability during the full eight months of monitoring (Sect. 2), although it can not be fully excluded yet. Moreover, from microlensing simulations, we expect that =1.2-2.4% if (Fig. 11), based on constraints on the source structure and mass function of compact objects found from the VLA 8.5-GHz light curves (Sects. 4, 5 and 7). This range agrees remarkably well with the observed modulation index =1.2% at 21 cm.

Supplementary to this argument, the difference in modulation-index between the lensed images would, in the case of scintillation, argue for either a very different Galactic ionized ISM (3.1; Sect. 3.1-2) towards the lensed images or a different image size (1.75; Sect. 3.2), although the latter might result from scatter-broadening. Furthermore, the longer variability time scales at 8.5 GHz (1 day; Figs. 1-2) are also difficult to explain in terms of scintillation, as well as the absence of variability with short time scales in several 12 h WSRT observations at 5 GHz (Koopmans et al. in prep.).

However, the strongest argument against scintillation remains the dominant presence of short-term external variability at 8.5 GHz in 1998, combined with the fact that in 1999 significant short-term variability is seen at 5 GHz, but almost none at 1.4 GHz.

Under the microlensing hypothesis, we find a consistent, although not unique set of jet-component parameters. A core plus a single-jet-component with a size of 2-5 µas, containing 5-11% of the flux density and moving superluminally with 926, can explain the modulation-index and variability time scale in both lensed images (Sects. 4-5). For image A we find a significantly higher average mass of compact objects (0.5 ), compared with those near image B. A much lower mass of compact object would result in a finer magnification pattern and thus in less variability. If image B is scatter-broadened, its microlensing modulation-index is reduced, which might change the lower-limit on the compact object mass.

If one, based on the evidence gathered thus far, accepts that the 1.4, 5 and 8.5-GHz short-term variability in B1600+434-A and B is dominated by microlensing, the profound consequence is that the dark-matter halo at 6 kpc above the plane of the disk-galaxy lens in B1600+434 is partly filled with massive compact objects. New WSRT, VLA and VLBI multi-frequency data is being obtained at the moment, which combined with a more comprehensive statistical analysis should provide us with refined constraints on the mass function of compact objects and the source structure (Koopmans et al. in prep.).

© European Southern Observatory (ESO) 2000

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