2005A&A...433..479K

We discuss the correlation between the evolution of the TeV emission and X-ray radiation observed in high-energy peaked BL Lac objects. We describe such a correlation by a simple power law F_TeV(t)∝F^x_X–ray(t). In the first part of this work we present correlations obtained for the activity of Mrk 501 observed in 1997 April and for the activity of Mrk 421 observed in 2000 February. Our results obtained for Mrk 501 show that the index of the correlation (x) may strongly depend on the width and position of the spectral bands used for the comparison. The result of the correlation which we have obtained for Mrk 421 is not informative. However, we discuss results of similar correlation obtained for this source by other authors. They report an almost quadratic (x∼2) correlations observed between the evolution of the TeV and X-ray emission. In the second part of this paper we present a phenomenological model which describes the evolution of the synchrotron and inverse Compton emission of a simple spherical homogeneous source. Neglecting the radiative cooling of the particles we derive analytical expressions that describe the evolution. Then we use a numerical code to investigate the impact of radiative cooling on the evolution. We show that different forms of correlations can be obtained depending on the assumed evolution scenario and the spectral bands used for the calculation. However, the quadratic correlation observed during the decay phase of the flare observed in Mrk 421 on 2001 March 19 appears problematic for this basic modeling. The quadratic correlation can be explained only for specific choices of the spectral bands used for the calculation. Therefore, looking for more robust solutions, we investigate the evolution of the emission generated by a cylindrical source. However this model does not provide robust solutions for the problem of a quadratic correlation. In principle the problem could be solved by the TeV emission generated by the self Compton scattering in the Thomson limit. However, we show that such a process requires unacceptably large values of the Doppler factor. Finally we briefly discuss the possible influence of the light travel time effect on our results.