2. Observations of VHE gamma rays
Observations were made with the University of Durham Mark 6 imaging gamma ray telescope operating at Narrabri NSW, Australia. The telescope is described elsewhere (Armstrong et al. 1999). It employs a 109-element imaging camera to discriminate gamma rays from hadrons and three flux collectors on a single mount to obtain an energy threshold of about 300 GeV. The telescope was operational from 1995 to 1999.
The results of initial observations of Cen X-3 in 1997 have been reported previously (Chadwick et al. 1998). Our complete Cen X-3 dataset comprises data from 31 hrs of observation (with an equal quantity of off-source observations) during 23 exposures in 1997 March and June (JD 2 450 508 - JD 2 450 606), 1998 March and April (JD 2 450 899 - JD 2 450 932) and 1999 February (JD 2 451 220 - JD 2 4512 30). An observing log is given in Table 1.
Table 1. Observing log for observations of Centaurus X-3 made with the University of Durham Mark 6 telescope at Narrabri, NSW, Australia. The numbers of ON and OFF source counts are the number of events surviving after data cleaning (see text).
Data were routinely taken in 15-minute segments. Off-source control observations were taken by alternately observing regions of sky which differ by minutes in RA from the position of Cen X-3 to ensure that on- and off-source segments have identical zenith and azimuth profiles and cosmic ray background response. The choice of alternate off-source segments which precede and follow the on-source segment allows for any small residual secular effects.
The data analysis employed follows the methods established in our previous detections of VHE emission from of Cen X-3 (Chadwick et al. 1998), PKS 2155-304 (Chadwick et al. 1999a) and the determination of limits from a range of extragalactic objects (Chadwick et al. 1999b) and pulsars (Chadwick et al. 2000). Data were accepted for analysis only if the sky was clear and stable and the gross counting rates in each on-off segment pair were consistent at the level. In our later observations, the sky clarity was assessed using an infra-red radiometer mounted paraxially with the telescope (Buckley et al. 1999).
Events were considered suitable for further analysis ("cleaned") if they were confined to within of the centre of the camera, and that they had sufficient information for reliable data analysis. For our analysis of the Cen X-3 dataset, a slightly larger size cut than has been employed in our PKS 2155-304 analysis has been used (800 digital counts rather than 500). This reflects the noisier images caused by the bright source field close to Cen X-3 and by the consequent running of the camera at a reduced HT setting. The number of ON and OFF source events remaining after this cleaning process is shown in Table 1.
Gamma ray candidates were selected using similar criteria to those employed in our detection of VHE gamma rays from PKS 2155-304. Gamma rays are selected through the shape of their images, using the criteria shown in Table 2. The parameters are as defined in Chadwick et al. (1999a). Slight differences in the values of the parameters used reflect the different zenith angle distributions and the effects of sky brightness. The final selection was based on the value of the pointing angle ALPHA , where a value less than was an indication of a gamma ray candidate.
Table 2. The image parameter selections applied to the analysis of data from the Cen X-3 data.
Our earlier report (Chadwick et al. 1998) was based on data recorded in 1997 March and June (JD 2 450 508 - JD 2 450 606) only. Assuming a collection area of and that our selection procedure retained of the original gamma ray events, the time averaged flux was estimated to be for GeV. Ongoing simulations suggest that our current selection procedure retains of the gamma rays. On this basis, the flux for the 1997 March and June (JD 2 450 508 - JD 2 450 606) data would be . The additional data taken in 1998 and 1999 provide fewer gamma ray candidates suggesting weaker TeV emission. In total, there were 617 excess events identified as gamma rays detected in 108 360 seconds of ON-source observation. Assuming, as above, a collection area of and that the current selection procedure retains of the original gamma ray events this yields a time averaged flux of at an average gamma ray threshold of GeV, defined as the energy at which the triggering probability was . The significance of the detection based on the total dataset is .
© European Southern Observatory (ESO) 2000
Online publication: December 15, 2000