Investigating the contribution of electron-positron pair production to accretion columns in bright X-ray pulsars

Until a decade ago, ultraluminous X-ray sources (ULXs) were believed to be powered solely by accreting black holes, whose large masses naturally explain their super-Eddington luminosities. This picture was challenged in 2014, when Bachetti et al. first detected pulsed emission from the ULX M82 X-2. Further detections led to a consensus within the community that the pulsations, together with other peculiar features observed, could be attributed to the presence of an accreting neutron star. The physics of accretion onto these compact objects is extremely complex, and strongly shaped by their ultra-strong magnetic fields, which govern the geometry of the flow, channeling the plasma onto the magnetic poles into a hollow structure known as the accretion column, and reduce the scattering opacity, thus raising the maximum luminosity the star can sustain. The presence of super-critical fields can also enable quantum electrodynamics (QED) processes that would otherwise be forbidden. The aim of this work is to determine whether the reduction of the scattering opacity induced by such fields remains a viable explanation for the extreme luminosities of pulsating ULXs once pair creation is consistently included in the description of the propagation of photons within the accretion column. Using an implementation based on the model of Mushtukov et al., we show that, at the equilibrium, pairs build up a persistent scattering target throughout the column, increasing the local opacity and representing a real obstacle to the free escape of radiation. Under this assumption, the magnetic reduction of the opacity alone cannot account for the observed luminosities exceeding 10^40 erg/s, and alternative mechanisms may be considered. However, the internal structure of the column is poorly known and equilibrium may not be reached everywhere; in that case, the luminosity reduction is modest and the magnetic explanation remains viable.