Astron. Astrophys. 361, 1073-1078 (2000)

2. Observations of the Crab Nebula with the HEGRA CT system

The HEGRA system of imaging atmospheric Cherenkov telescopes is located on the Canary Island of La Palma, on the site of the Observatorio del Roque de los Muchachos. The telecope system consists of five telescopes, with a mirror area of 8.5 m² and a focal length of 5 m. A sixth prototype telescope is operated in stand-alone mode and is not used for the analysis presented here. The system telescopes are arranged at the corners and in the center of a square of about 100 m side length. The alt-azimuth mounted telescopes are equipped with cameras consisting of 271 photomultipliers (PMTs). Each PMT views an area of the sky of diameter; the field of view of each camera is about . Cherenkov images of air showers are recorded whenever two telescopes trigger simultaneously; the trigger condition requires that two neighboring PMTs exhibit signals equivalent to 8 or more photoelectrons. Typical trigger rates are around 15 Hz, for an energy threshold of 500 GeV for vertical gamma rays. Details about the HEGRA IACTs and their performance can be found in Daum et al. (1997), Aharonian et al. (1999b), Hermann (1995), Bulian et al. (1998).

The Crab Nebula was observed in each season since the HEGRA IACT system commenced operation in late 1996, initially with three telescopes, later with four and since late 1998 with the complete set of five telescopes. For this analysis, only data taken in the years 1997 and 1998 were used, acquired with at least four telescopes under good weather conditions; in order to be able to compare with earlier Mrk 501 data taken with four telescopes, data from the fifth telescope was not used in the most recent five-telescope data sets. The quality-selected data set amounts to an integral observation time of about 155 h, and includes about 6.3 million events. For the final analysis, only data taken at zenith angles of less than were included, with about 3.5 million events remaining. The Crab Nebula was observed in the so-called wobble mode, with the source offset by in declination relative to the telescope axes. The sign of the offset alternated every 20 min. A region offset by the same amount, but in the opposite direction, is used for background estimates, avoiding the need for special off-source observations. Since the stereoscopic reconstruction of air showers provides an angular resolution of typically 6´, signal and background regions are well separated.

The techniques for data analysis are similar to those documented, e.g., in Aharonian et al. (1999b, 1999d). Direction and impact point of an air shower are reconstructed from the stereoscopic views of the shower. Based on the measured impact point and the known (Aharonian et al. 1999a) distribution of Cherenkov light as a function of the distance to the shower axis, an energy estimate is derived, with a typical resolution of 20%. Gamma-ray candidates are selected on the basis of image shapes. Given the core distance and the intensity (the Hillas size parameter (Hillas 1985)), the expected width of a gamma ray image is determined. The measured width values are normalized to this value, and averaged over telescopes. A cut on the resulting mean scaled width 1.2 retains most gamma-rays, but rejects the bulk of the cosmic-ray showers.

In detail, the reconstruction of shower geometry differs somewhat from the techniques used so far. Whereas the normal reconstruction procedure combines images from all telescopes regardless of their quality, the new procedure assigns - on the basis of the Monte Carlo simulations of Konopelko et al. (1999b) - errors to the relevant image parameters (the location of the image centroid and the orientation of the image axis). These errors depend on the intensity and the shape of the images and are propagated through the geometrical reconstruction, resulting in error estimates (or, to be precise, a covariance matrix) for the shower parameters. Details of the algorithm are given in Hofmann et al. (1999). Depending on the characteristics of an event, an angular resolution between 2´ and more then 10´ is predicted, with average value slightly below 6´ ¹. The ability to select subsets of events with better-than-average resolution will be used extensively in the analysis of the size of the VHE emission region in the Crab Nebula.

© European Southern Observatory (ESO) 2000

Online publication: October 10, 2000
helpdesk.link@springer.de