Orateur
Description
Radio pulsars exhibit frequency-dependent variations in the shape of their pulse profiles that introduce systematic delays in the pulse times of arrival (TOA) across an observing band. Some Pulsar Timing Arrays model these delays using a frequency-dependent (FD) log-polynomial function whose coefficients are fit to TOA measurements. However, these FD parameters are prone to absorbing power from other chromatic effects, such as interstellar dispersion and scattering, leading to models that overestimate pulse-profile evolution and are not self-consistent across datasets. To better decouple these effects, we introduce two new mathematical formulations of frequency-dependent profile evolution, based on monomial and Legendre polynomial expansions, and test them using both observational data and targeted simulations. We find that the Legendre-based model generally predicts delays that more closely match the observed frequency-resolved pulse profiles, are more self-consistent across datasets, and better recover the injected profile evolution in simulated data.