1 Introduction

Thanks to $\gamma$-ray observations of EGRET, on board CGRO, we now know the overall spectral energy distribution (SED) of blazars. They are characterized, in $\nu$- $\nu F(\nu)$ plots, by two broad peaks. It is believed that the first, located in the IR-soft X-ray band, is due to synchrotron emission, while the second is due to the inverse Compton process by the same electrons producing the synchrotron part of the spectrum (Maraschi 1992; Sikora et al. 1994; but see Mannheim 1993 for a different interpretation). The 0.1-10 keV emission of blazars is therefore located in the minimum between the two peaks, where both processes (synchrotron and inverse Compton) can contribute. Observations in this band are therefore useful to characterize the relative importance of both processes. This can constrain models, allowing a better determination of the location of both peaks. Usually, a steep power law energy distribution in the X-ray band (with a spectral energy index $\alpha>1$, with $F \propto \nu^{-\alpha}$) is due to the tail of the synchrotron spectrum, while $\alpha<1$ flags the dominance of the inverse Compton spectrum. There are exceptions to this rule, such as HBL (High energy peaked BL Lacs) in a flaring state, which show a synchrotron spectrum peaking above 10 keV. One dramatic example of this behavior is Mkn 501, whose synchrotron peak energy, usually located below/at ${\simeq}1$ keV, shifted to 100 keV or more during its flare in April 1997 (Pian et al. 1998). In these cases the X-ray spectrum, usually steep during quiescence, becomes much flatter during flares. Fossati et al. (1998) (F98 hereafter) have shown that blazars form a sequence, with their SED changing in a continuous way as their bolometric power changes: low luminosity objects (HBL) have the synchrotron peak in the UV-soft X-ray band, and the inverse Compton peak between the GeV and the TeV band. The two components have approximately the same power. As the bolometric luminosity increases, both peaks shift to lower frequencies, and the Compton peak becomes increasingly dominant. This trend offers the opportunity to unify in a single scheme the many flavors of existing blazars, and calls for a physical explanation (see e.g. Ghisellini et al. 1998).

To check the reliability of this trend we have collected data for all blazars having available spectral information in the X-ray band. For the soft band [0.1-2 keV], most of the results come from ROSAT, while for the 2-10 keV band the results are gathered from the EXOSAT, ASCA and BeppoSAX satellites.

Besides the data already published, we searched for unpublished data in the ASCA public archive, finding 12 observations of 7 sources. Results of the analysis of these data are presented here. We then add these sources to our sample.

The entire sample forms the largest database in the X-ray range: 421 spectra of 268 blazars. The X-ray data have been complemented by additional information regarding the redshift (when available), the radio flux at 5 GHz and the optical flux (in the V band).

The paper is organized as follows; Sect. 2 is devoted to the analysis of the 12 ASCA spectra, while in Sect. 3 we present the entire set of data. In Sect. 4 we compare the results in the soft and the hard X-ray bands and check for correlations with other spectral parameters, such as the broad band spectral indices connecting the radio with the optical fluxes, the optical with the X-ray fluxes, and the radio with the X-ray fluxes. In Sect. 5 we discuss our findings in the framework of the scenario proposed by Fossati et al. (1998), suggesting an improvement connected to a possible physical difference between low and high power sources.