In gamma-ray binaries neutron star is orbiting a companion that produces a strong stellar wind. We demonstrate that observed properties of ‘stellar wind’–‘pulsar wind’ interaction depend both on the overall wind thrust ratio, as well as more subtle geometrical factors: the relative direction of the pulsar’s spin, the plane of the orbit, the direction of motion, and the instantaneous line of sight. Using fully 3D relativistic magnetohydrodynamical simulations we find that the resulting intrinsic morphologies can be significantly orbital phase-dependent: a given system may change from tailward-open to tailward-closed shapes. As a result, the region of unshocked pulsar wind can change by an order of magnitude over a quarter of the orbit. We calculate radiation maps and synthetic light curves for synchrotron (X-ray) and inverse-Compton emission (GeV-TeV), taking into account $\gamma $–$\gamma $ absorption. Our modelled light curves are in agreement with the phase-dependent observed light curves of LS5039.

Gamma-ray observations by the Fermi Large Area Telescope (LAT) have been used very successfully in the last 9 years to detect more than 200 gamma-ray pulsars. Sixty of these have been found by directly searching for pulsations in the gamma-ray data, but only one binary MSP has been found this way. Pulsars in binaries are often difficult to detect in radio data because of large eclipses, and some binary MSPs may even be radio quiet. For those, a gamma-ray blind search might be the only possibility for detection. While searches for isolated pulsars up to kilohertz frequencies are already computationally very challenging, blind searches for binary gamma-ray pulsars are simply infeasible without further knowledge of their orbital parameters. Here we present methods with which we can conduct searches for candidate binary gamma-ray pulsars for which orbital constraints are known from optical observations of a likely companion star. We also highlight some example sources where these methods have been used.