Since the detection of this source at TeV energies (Punch et al. 1992), and the observation of correlation in the flaring activity at X-ray and TeV energies, observations of Mkn 421 at X-rays have become more important. The temporal coincidence of flaring in X-ray and TeV supports emission models where both components are produced by the same population of electrons with the X-rays at the endpoint of the Synchrotron spectrum and the -rays at the endpoint of an inverse Compton spectrum (Takahashi et al. 1996). This is consistent with the overall picture drawn from the SED of Mkn 421 which is characterized by two maxima, one in the UV/soft X-ray band, and the second at GeV to TeV energies.
There is strong evidence that the X-ray spectrum of High-energy peaked BL Lac objects (HBL) is concave (Sambruna et al. 1994; Tashiro et al. 1995; Takahashi et al. 1996; Sambruna et al. 1997). However, even the "standard" broken power-law model is unable to reproduce the spectral curvature of Mkn 421 observed here. A system of two absorption edges, whose threshold energies are broadly consistent with highly ionized Neon species NeIX and NeX , is needed to get an acceptable fit quality. However, if the edges represent the imprinting of a photoionized absorber along the line of sight, it is difficult to explain the presence of highly ionized Neon, while no Oxygen feature is detected. Assuming the K-shell photoionization cross section formulae from Band et al. (1990), the equivalent Neon column density inferred by the measured optical depth is cm-2, assuming that the detected species represent the bulk of the elemental abundance. In a photoionized plasma with spectral index , the simultaneous presence of Ne IX and Ne X implies an ionization parameter , for which OVIII should be the dominant Oxygen ionic species (see e.g. Kallman & McCray 1982). Steeper spectra should only enhance the relative Oxygen/Neon ratio. The observational upper limit on Oxygen abundance is, however, more than one order of magnitude lower than Neon ( cm-2). This scenario is viable only if deep Oxygen edges were detected as well. However, ROSAT found no evidence for these spectral features (Fink et al. 1991). We conclude that such a description of the observed spectrum is unphysical and can be discarded.
A good fit is obtained with a model where the spectral steepening occurs gradually, hence no additional spectral component is required. In this framework, the hardening of the spectrum with increasing flux is basically due to a change of the X-ray spectral properties above 4 keV.
It is generally believed that the radiation from BL Lac objects is relativistically beamed (Blanford & Rees 1978) and that the beam points directly towards the observer (Marscher 1980; Königl 1981). The shape of the spectrum and the measurement of polarization in the radio to optical bands provide the basis for the interpretation of the emission in terms of SSC models. Here emission up to X-ray wavelengths is caused by a relativistic population of electrons via the synchrotron process, while the -ray photons are created by inverse Compton scattering of the same electron population with the ambient photons (Bregman et al. 1990; Kawai et al. 1991). In the SSC model, the radiation output is driven by the highest energy electrons. The observed curvature of the X-ray spectrum can be explained by assuming that the jet is structured and the electron distribution is located in a smaller (inner) region of the jet with increasing energy. An increase in the synchrotron flux means an extension of the region where the high energy electrons are injected, probably in connection with increased activity of the nuclear engine. Alternatively, the hardening of the spectrum with increasing X-ray flux could be due to the injection of new electrons in the jet, which are highly energetic and therefore produce a flatter spectrum before they suffer significant radiative losses.
Einstein observations of a small sample of X-ray bright BL Lac objects indicate that an absorption feature at 0.65 keV may be a ubiquitous feature in their soft X-ray spectra (Majedski et al. 1991). This feature was first claimed to be detected by Canizares & Kruper (1984) in PKS 2155-304. It was interpreted as resonance absorption in OVIII ( eV). However, ROSAT observations (Fink et al. 1991) of Mkn 421 found no evidence for such a spectral feature. The BeppoSAX observation of PKS 2155-304 did not detect such a feature either (Giommi et al. 1998), casting further doubts on its existence in BL Lac objects. In agreement with these findings, our data show no evidence for any spectral features in the spectrum of Mkn 421. In particular no absorption feature is observed around 0.6 keV, the 90% upper limit on the equivalent width of a resonance absorption in OVIII being only 10 eV.
© European Southern Observatory (ESO) 1999
Online publication: December 22, 1998