-ray emission
The strong radio source PKS 1830-211 was first proposed to be a
gravitational lensed QSO by Pramesh Rao & Subrahmanyan (1988). High resolution
radio images obtained from several interferometric arrays have revealed that the
source has a ring-like structure with two bright components on sub-arcsecond
scales (Jauncey et al. 1991). This suggests a close alignement
of the lensed source behind the lensing object. Actually, two absorption
systems have been detected at 
(Wiklind & Combes 1996) and
(Lovell et al. 1996), so it seems likely that
the image of the background QSO (with a redshift 
)
is lensed by two different extragalactic objects (probably gas-rich
spirals). The background source can be modeled as a core-knot-jet
structure, similar to other flat-spectrum QSOs which are known to be
-ray emitters (Nair et al. 1993). The -ray
spectrum of 2EG 
is remarkably similar to several spectra of
QSOs detected by EGRET, like 0234+285 and 0454-463 (von Montigny et al.
1995). These high-energy spectra are much steeper than those
expected for galactic sources like pulsars. This fact, along with the
spatial coincidence, strongly suggests the identification of 2EG with PKS 1830-211.
The presence of variability in the time history of 2EG
could provide additional support to the proposed identification. This time
history is presented in graphical form in Fig. 4 (click here) for a 
yr time span. We have used the data from the second EGRET catalog corrected
and completed by McLaughlin et al. (1996). Systematic errors
over the statistical uncertainties of EGRET flux measurements are difficult
to estimate. These errors can be due to uncertainties in the instrumental
calibration as a function of energy, uncertainties in angle within the
instrument, and errors in the galactic diffuse radiation model. McLaughlin
et al. (1996) have quantified these systematic errors assuming that
-ray pulsars are nonvariable sources, obtaining an additional
uncertainty of 6.5% 
1.0% which is included in the error bars in
Fig. 4 (click here).
Figure 4: Time history of 2EG over a period of
3.5 yr
A 
-variability analysis of the entire light curve (see Romero et al.
1994) shows that the source behaviour is "probably variable''. This
is mainly due to the large errors in the flux values during the viewing
periods from mid-1992 to mid-1993. However, if we restrict our analysis to
the lapse August 1991 - February 1992 we find clear evidence for
significant variability. The flux increased from 
cm-2 s-1 in August 1991 to 
cm-2
s-1 at the end of 1991, and then decreased to 
cm-2 s-1 in February 1992. This implies a flux change by a factor
of at least 2.5 over a time scale of 6 months. This behaviour seems
to be incompatible with a pulsar (see Ramanamurthy et al.
1995) or SNR-type source. Conversely, rapid -ray variability
in PKS 1830-211 could be produced both intrinsically or by gravitational
microlensing. In this latter case, the background -ray region in
the innermost part of the QSO is magnified by a compact massive object in
one of the intervening galaxies. The variability time scale is given by the
time spent by the line of sight to the source in crossing the microlens
Einstein radius, i.e. 
where 
is the microlens Einstein radius, 
is a distance in Gpc obtained from the source-lens, lens,
and source angular-diameter distances in a Robertson-Walker Universe, M is the
mass of the lens in units of solar masses, and v3 is the velocity V of the
lens in units of 103 km s-1 (see Romero et al. 1995 and
references therein for details). Assuming a redshift 
for the
background source and 
for the microlens, we find
that 
if 
(we have considered H0=100
km s-1 and q0=1/2). Consequently, a MACHO-like object in the halo
of the foreground galaxy could produce the observed variability.
Flux variations will occur in this scenario just if the angular radius of the
source in the lens plane is smaller than the Einstein angular radius of the
microlens. This imposes the constraint that the size of the -ray
emitting region should be 
cm 
pc, in good accordance with the sizes expected for the -spheres in
blazars (e.g. Blandford & Levinson 1995). Since the
-spheres are much smaller than the compact radio cores, no
correlation with lower frequency variability should be expected for the
-ray microlensing events. In fact, the sizes of the optical and
radio emitting regions in the lens plane should largely exceed the Einstein
ring sizes for small compact objects and, consequently, no significant
amplifications of the lensed images should happen at these wavelengths. On
the other hand, intrinsic -ray variability seems to occur in the
initial phases of high radio outbursts (e.g. Valtaoja & Teräsranta
1995). This fact could be used to discriminate between future
intrinsic and extrinsic -ray variability events in PKS 1830-211.