Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 359, 337-346 (2000)

Previous Section Next Section Title Page Table of Contents

4. Results

Table 3 presents a summary of data accepted for analysis from the 1994 and 1995 observing seasons.


Table 3. Summary of 1994/1995 W28 CANGAROO data accepted for analysis. The normalisation factor [FORMULA] used in adjusting the statistical significance of the post-cut ON-OFF excess, and the raw trigger efficiencies for gamma-rays above 0.8 TeV ([FORMULA]) are included.

The skymaps of significance S (Fig. 4) for both years' data, do not indicate any significant point-like excesses over a [FORMULA] search for [FORMULA]-ray emission. Table 4 summarises the 3[FORMULA] upper limits for the positions of a number of interesting sites within the W28 region. It should be pointed out that our results will have a systematic error of the order [FORMULA]%, based on uncertainties in the trigger conditions, mirror reflectivity and spectral index adopted in the simulations (Yoshikoshi 1996). For example, in the next section we compare our results to a model producing a gamma-ray flux with spectral index of [FORMULA], harder than the -1.6 spectrum used in simulations. Such a difference however will contribute a systematic of [FORMULA]5%.

[FIGURE] Fig. 4. SDF skymaps of ON-OFF significance for W28 1994 data (left), 1995 data (right). The skymaps extend [FORMULA] from the respective tracking centres (indicated by the circled cross) of each year's data. Contour levels indicate the same ON-OFF excess over 10 intervals between the maximum and minimum excess values of each skymap.


Table 4. Details of various sites considered as potential emitters of TeV gamma-rays in the W28 region for CANGAROO 1994 and 1995 data and corresponding 3[FORMULA] flux upper limits. The statistical significance S is calculated from Eq. (1) with the normalisation factor [FORMULA] given by Table 3.
a) Radio point source (Andrews et al. (1983)
b) Extended source of radius 0.25o encompassing the molecular clouds.
c) Point source at pulsar position given by from Kaspi et al. 1993).
d) Point source at average position of 1720 MHz OH maser sites E and F as defined by Claussen et al. (1997). See text.
e) Highest pointlike significance within 95% error circle (0.32o radius) of EGRET source. Quoted position is that for the error circle centre (Hartman et al. 1999). See also f) below.
f) Positions of each pointlike maximum significance within the 3EG error circle.

The following positions were considered as potential sources; the position A83 given by Andrews et al. (1983), the pulsar PSR J1801-23, the two strongest (by an order of magnitude) masers, labelled E & F by Claussen et al. (1997), and the EGRET source 3EG J1800-2338. In addition, an extended region of radius 0.25o centred on a position to encompass the molecular clouds was considered. An average position was used for the masers E and F since they are separated by only 0.02o. In assuming that the EGRET source was point-like, a search for the highest point significance within the 95% error circle was carried out. Since the statistical degrees of freedom for a non-a priori search for point-like emission over the skymap is [FORMULA]100, the highest ON-OFF excess within the EGRET error circle must be interpreted with a similar statistical penalty in mind.

The point spread function (PSF) for a pure gamma-ray signal can be used to assess the location accuracy of the proposed source positions. Monte Carlo simulations show that the PSF for gamma-rays increases slightly from a standard deviation of 0.2o on-axis to 0.22o for sources at the skymap corners. The positions of our candidate sources (Table 4) lie within 1.0o on-axis. Taking a conservative estimate of the PSF as 0.22o, a simplistic estimate of the source location error is obtained by adding in quadrature 0.22/[FORMULA]=0.02o to the the absolute tracking precision (0.02o), giving [FORMULA]. We have used a value of 100 which is representative of the 3[FORMULA] upper limit excesses (N in Eq. (4)) calculated here. We can therefore say that the features listed in Table 4 would be resolved by the CANGAROO 3.8m telescope if they were point sources of TeV gamma rays. From Table 4 we can also see that the positions of the highest significance within the EGRET 95% error circle for both year's data are also not co-located, with their separation, [FORMULA], easily exceeding the estimated PSF.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: June 30, 2000