Context: Accretion-powered X-ray pulsars show highly energy-dependent and complex pulse-profile morphologies. Significant deviations from the average pulse profile can appear, in particular close to the cyclotron line energies. These deviations can b
e described as energy-dependent phase lags, that is, as energy-dependent shifts of main features in the pulse profile. Aims: Using a numerical study we explore the effect of cyclotron resonant scattering on observable, energy-resolved pulse profiles. Methods: We generated the observable emission as a function of spin phase, using Monte Carlo simulations for cyclotron resonant scattering and a numerical ray-tracing routine accounting for general relativistic light-bending effects on the intrinsic emission from the accretion columns. Results: We find strong changes in the pulse profile coincident with the cyclotron line energies. Features in the pulse profile vary strongly with respect to the average pulse profile with the observing geometry and shift and smear out in energy additionally when assuming a non-static plasma. Conclusions: We demonstrate how phase lags at the cyclotron energies arise as a consequence of the effects of angular redistribution of X-rays by cyclotron resonance scattering in a strong magnetic field combined with relativistic effects. We also show that phase lags are strongly dependent on the accretion geometry. These intrinsic effects will in principle allow us to constrain a systems accretion geometry.
(abbreviated version of the abstract) We study the physics of cyclotron line formation in the high-energy spectra of accreting X-ray pulsars using Monte Carlo methods, assuming that the line-forming region is a low-density electron plasma in a sub-cr
itical magnetic field. We investigate the dependence of the shape of the fundamental line on angle, geometry, optical depth and temperature. We also discuss variations of the line ratios for non-uniform magnetic fields. These numerical predictions for the line profiles are linked to results from observational data analysis using an XSPEC model based on the Monte Carlo simulations. We apply this model to observational data from RXTE and INTEGRAL. The predicted strong emission wings of the fundamental cyclotron feature are not found in observational data, hinting at a bottom illuminated slab geometry for line formation.
