Astron. Astrophys. 349, 11-28 (1999)

1. Introduction

Mkn 501, an active galactic nucleus (AGN) at a redshift , was discovered several years ago as a faint source of TeV -radiation (Quinn et al. 1996; Bradbury et al. 1997). In 1997 it turned into a state of high activity, unique in both its strength and duration. The TeV emission of the source from March to September 1997 was characterized by a strongly variable flux. It was on average more than three times larger than the flux of the Crab Nebula, the strongest known persistent TeV source in the sky. Fortunately, the time period of the outburst coincided with the source visibility windows of several ground-based imaging atmospheric Cherenkov telescopes (IACTs) designed for the detection of very high energy (VHE) cosmic -rays. Thus almost continuous monitoring of Mkn 501 in TeV -rays with several IACTs (CAT, HEGRA, TACTIC, Telescope Array, Whipple) located in the Northern Hemisphere was possible (e.g. Protheroe et al. 1997).

The observations of Mkn 501 by the HEGRA stereoscopic IACT system during this long outburst made a detailed study of the temporal and spectral characteristics of the source possible, based on an unprecedented statistics of more than 38,000 TeV photons (Aharonian et al. 1999a; hereafter Paper 1). The "background-free" detection of -rays , with an average rate of several hundred -rays per hour (against background events caused by charged cosmic rays), allowed us to determine statistically significant signals for minute intervals during much of the 110 h observation time, spread over 6 months. Moreover, it was possible to monitor the energy spectrum of the source on a daily basis. Within the errors the energy spectrum maintained a constant form over the range from 1 TeV to 10 TeV. This was the case even though the flux varied strongly on time scales day. We believe that this is an important result, and it was to some extent unexpected.

The diurnal spectra exhibit a power-law shape at low energies (between 1 TeV and several TeV), with a gradual steepening towards higher energies (Paper 1). Such a spectral form could not be unequivocally ensured in the first analysis which was performed during the period of activity of the source, since the systematic errors of the recently commissioned stereoscopic system of HEGRA were not well studied at this time. As a consequence, the energy spectrum could not be determined more precisely than implied by a power law fit (Aharonian et al. 1997a), even though the tendency for a gradual steepening of the observed spectra was noticed (Aharonian et al. 1997b). The results of Paper 1 and the new results in a broader energy interval presented below are based on detailed systematic studies (see Paper 1 and Konopelko et al. 1999), and extend and supersede these previous results. This allowed us to come to the definite conclusion that the spectrum determined in the energy region from 1 to 10 TeV steepens significantly (Paper 1). A similar tendency has been found also by the Whipple (Samuelson et al. 1998), Telescope Array (Kajino et al. 1999, private communication), and CAT (Djannati-Atai et al. 1999) groups. Independent spectral measurements by the HEGRA telescopes CT1 and CT2 will be published elsewhere.

Apart from its astrophysical significance, the constancy of the spectral shape has the important practical consequence that it allows to measure the spectrum with small statistical errors also in the energy regions below 1 TeV and above 10 TeV. Indeed, the low photon statistics of the detector in both "extreme" energy bands (towards low energies basically due to the decrease of the detector's collection area; towards high energies due to the steep photon spectrum) can be drastically increased by using the data accumulated over the whole period of observations.

In Sect. 2 we describe the HEGRA stereoscopic system and the specific form of the data analysis, based on Monte Carlo simulations of both, the air showers and the detection system. The data sample is the same as in Paper 1 and is described in Sect. 3. A detailed study of the systematic errors in the spectrum derivation is contained in Sect. 4; most of this methodology was actually developed in the context of the Mkn 501 data analysis. We believe that this is the first study of this kind. The experimental results are then presented in Sect. 5, whereas Sect. 6 attempts a first discussion. This discussion concentrates on the -ray results of Sect. 5 and what one can learn from them alone. A multiwavelength analysis is clearly outside the scope of this paper. Our conclusions are contained in Sect. 7.

Readers only interested in the astrophysical results, should skip Sects. 2-4 and proceed to Sect. 5.

© European Southern Observatory (ESO) 1999

Online publication: August 25, 1999
helpdesk.link@springer.de