Multi-wavelength variability studies have been used frequently to determine the physical conditions of the central engines of BL Lac objects. However, in most cases the irregularities in the variability patterns and the puzzling spectral behavior of these sources have frustrated simple interpretations.
The extended ROSAT observations of PKS 2155-304 in May 1996 and November 1996 suggest that the X-ray variability is due to several "modes": one, varying smoothly with time scales of more than a week, a second, consisting of individual shots with duration of one day and repetition rates of about one per day and, possibly, low-amplitude `flickering' with scales of 10 minutes. The sharp, isolated flares, such as seen in May 1994 with ASCA and in November 1996 /1997 with BeppoSAX have timescales similar to those of the "shots" and could represent high amplitude "events" within the same phenomenology. Unfortunately the structure function is not sufficiently well determined to reveal critical timescales. However it is important to recall that the latter flares show significant spectral variability while the low amplitude "shots" first revealed with ROSAT were essentially achromatic. While the absence of spectral variations could be due to the limited spectral band of ROSAT and to the limited amplitude of the events we cannot exclude that they represent an intrinsically different "mode" of variability.
The long term variations might be directly related to geometrical changes of the system, either to helical trajectories of moving knots in a relativistic jet (Camenzind & Krockenberger 1992) or to a precession of the jet. For example, the residual light curve of May 1996 is well fit by a straight line with a slope of 0.58 HRI counts/sec/day. If we interpret this smooth increase of flux in terms of a varying Doppler boosting caused by changes of the viewing conditions of a precessing relativistic jet we obtain a bulk velocity of the flow of (i.e. ) and a viewing angle in the range .
The short-term variability, the shots, occuring on time-scales of about a day can be successfully explained either in the accelerating inner jet model by a sudden increase of the density of energetic electrons (Georganopoulos & Marscher 1997) or by particle acceleration at shocks traveling down the relativistic jet (Kirk et al. 1998). As the HRI detector does not provide energy information we cannot use any spectral information for a more detailed comparison of the different radiation models (see Urry et al. 1997).
In the framework of a time dependent model of synchrotron self-Compton emission from a homogeneous region the quasi-symmetric shape of the short flares can be reproduced quite well, provided that light-travel effects are properly taken into account (Chiaberge & Ghisellini 1999, Kataoka et al. 1999). In this framework, the symmetric flares observed at optical and X-ray wavelengths at several occasions, strongly constrain the injection and the cooling time scales: the injection time must be of the same order as the source light-crossing time scale , while the cooling time must be shorter than . With such a model Kataoka et al. (1999) were able to reproduce the temporal evolution of the spectrum as well as the light curve during the flare of PKS 2155-304 in May 1994, after determining all input parameters (required to specify the model) in a self-consistent way from observables obtained during the quiescence state of the source.
Alternatively, one might attribute the low amplitude shots (which seem to repeat, but not periodically) to geometry and the long term increase and large flares to particle acceleration. The rotation period expected for a magnetic helix would be 2R/c, i.e., about 2 hours for and . It could obviously be stretched to 12 hours to explain the recurrent (supposedly achromatic) fluctuations. Future observations with good spectral resolution and sufficient temporal coverage might allow to discriminate these scenarios.
The systematic pattern in this variability, most clearly noticed in the striking similarity between light curves taken five years apart, indicates the presence of some kind of temporal memory in the source which is not accounted for in the above models but could be accounted for by variations associated to a varying aspect like a helix associated with the rotational dragging of a magnetic field. If the shots or outbursts, characteristic for the short-term variability, occur in blobs of enhanced activity traveling down the jet then they cannot be generated randomly; a rather well defined physical mechanism must at least initiate the generation of these regions. Finally, geometrical viewing constraints might play a role as well for the observed variability pattern. These temporal structures further seem to firmly rule out a microlensing interpretation put forward to explain the nearly achromatic light curve of November 1991 (Treves et al. 1997).
Finally, the accumulation of all the individual ROSAT observations clearly shows that in general, long and continuous observations are required to unravel the physical conditions of AGN and PKS 2155-304 might be one of the prime objects where this can be achieved with current X-ray missions.
© European Southern Observatory (ESO) 2000
Online publication: October 30, 19100