2 Data analysis and results

a) Coincidences with BATSE events

We considered 292 GRBs observed by BATSE [(Brainerd 1998)] during $1992 \div 1998$ occurred in the field of view of EASTOP with a zenith angle $\theta \leq 60^{\circ}$. For each event the measured number of counts C during the $\Delta t_{90}$ time interval in which BATSE detected 90% of the flux is compared with the average number of counts B expected from the background (evaluated in 600 s around the burst).

No significant excess has been observed and the distribution of the excesses C-B is fully explained by poissonian fluctuations (see Fig. 1). Looking for possible delayed or anticipated excesses with respect to the BATSE event, the same analysis procedure has been performed in a 2 hours interval centered on the BATSE time, changing the width of the time window from 1 to 200 s. Also in this case no excess was found.

  

Figure 2: Upper limits on the $10~{\rm GeV}\div 1~{\rm TeV}$ energy fluence obtained for 292 BATSE events, as a function of the zenith angle. Full squares correspond to 4 GRBs observed by Beppo-Sax (GRB 971214, GRB 970508, GRB 971227 and GRB 980703)

Figure 2 shows the obtained upper limits on the energy fluence in the range $10~{\rm GeV}\div 1~{\rm TeV}$ for the 292 bursts analyzed, in the time window $\Delta t_{90}$,as a function of the zenith angle. The fluence upper limits have been calculated at 5 standard deviations level, assuming a GRB spectrum dN/dE $\propto\ E^{\alpha}$ with $\alpha=-2$, extending up to 1 TeV.

The gamma-ray absorption in the intergalactic space through $\gamma + \gamma \rightarrow {\rm e}^+ {\rm e}^-$ pair production affects the high energy part of the spectra of GRBs located at cosmological distances. According to [(Salomon&Stecker)] the gamma-ray flux from a source at redshift z=1 is reduced by a factor $\sim 3$ at E=50 GeV and by a factor $\sim 30$ at E=100 GeV. Assuming the GRB sources at z=0.5(1.0), the obtained fluence upper limits have to be increased by a factor $\sim 8(16)$.

b) All sky survey

GRBs can be searched as short duration increases ($\Delta t \leq 1$ s) in the flux of secondary charged particles. In each second, the counting rate C is compared with the expected background B calculated averaging the counting rate in 15 minutes around. Figure 3 shows the distribution of the differences ${\rm C}-{\rm B}$ in unit of standard deviations, obtained assuming poissonian fluctuations, for a total time of measurement of 3.6 10⁴ hours. The data are well fitted by a Gauss distribution with $\sigma=1.17$,showing the stability and good performance of the detector over a long time exposure.

A single statistical significant excess (10.6 standard deviation) has been observed on 1992 July 15 at 13:22:26 UT [(Aglietta 1993)]. Assuming this excess due to a gamma-ray burst with a zenith angle $\theta=30^{\circ}$, the correspondent $10~{\rm GeV}\div 1~{\rm TeV}$ energy fluence is $1.5~10^{-3}~{\rm erg}~{\rm cm}^{-2}$.

  

Figure 3: Distribution of the excesses of duration $\Delta t=1$ s in units of standard deviations, obtained in the "all sky" survey