We considered the cascade initiated by photons or electrons injected from the compact object in the radiation field of the massive companion in the Cen X-3 system, assuming that the secondary electrons are isotropised by the magnetic field in the binary. It is found that the features of the escaping photons from such massive binaries (the light curves, photon spectra as a function of the phase of the compact object) may allow the determination of which particles are injected by the compact object, i.e. photons or electrons. If the cascades are initiated by electrons then the escaping secondary cascade photons should not show features of modulation with the pulsar period. This seems to be in contradiction with the observations of Cen X-3 at energies above 100 MeV. Therefore we reject such a model. If primary photons are injected isotropically with a power law spectrum and spectral index -2, then the spectrum of escaping photons at energies below GeV is dominated, for most of the phases, by the primary non-cascading photons. Then, the modulation with the pulsar period can be observed. At TeV energies, the modulation of the photon flux with the 2.09 day binary period should be strong. The observation of modulation of the TeV signal with the orbital period of the system have been reported by earlier observations of Cen X-3 system (mensioned above) but not confirmed recently (Chadwick et al. 1999b). We think that this problem needs further investigation. The observations with the Mark 6 telescope had a tendency to cluster near orbital phases that had shown emission before, as commented in Chadwick et al. (1999b). Therefore the coverage of the orbit was not uniform. Also the presence of the `high' and `low' states observed in the system may have strong influence on the -ray light curve.
If the lack of modulation of the TeV photon flux with the Cen X-3 binary period is real, then a more complicated model has to be investigated. An extended source, e.g. a shock inside the binary system, injecting primary photons or electrons which initiate cascades in the soft radiation of a massive star should be developed. However such computations will require much more computing time in order to get satisfactory statistics, and therefore are left for future work.
In the present calculations we neglected the X-ray radiation field produced by the compact object. Its energy density erg s-1 (Giacconi et al. 1971) is comparable to the energy density of thermal photons from the massive star, so their photon density is a few orders of magnitude lower inside the volume of the binary system. We neglect also the heating effects of the massive star by the X-rays coming form the compact object since the power emitted by the stellar surface is higher than the X-ray power falling on the massive star from the orbital distance of the compact object equal to 1.4 stellar radii. However note that X-rays, produced in the accretion disk around a compact object, i.e. close to the production site of primary photons, may absorb -rays if they are also produced close to the inner disk radius (see e.g. Bednarek 1993).
© European Southern Observatory (ESO) 2000
Online publication: October 24, 2000