Is there a spectral turnover in the spin noise of millisecond pulsars?

Abstract: Pulsar timing arrays provide a unique means to detect nanohertz gravitational waves through long-term measurements of pulse arrival times from an ensemble of millisecond pulsars. After years of observations, some timing array pulsars have been shown to be dominated by low-frequency red noise, including spin noise that might be associated with pulsar rotational irregularities. The power spectral density of pulsar timing red noise is usually modeled with a power law or a power law with a turnover frequency below which the noise power spectrum plateaus. If there is a turnover in the spin noise of millisecond pulsars, residing within the observation band of current and/or future pulsar timing measurements, it may be easier than projected to resolve the gravitational-wave background from supermassive binary black holes. Additionally, the spectral turnover can provide valuable insights on neutron star physics. In the recent study by Melatos and Link, the authors provided a derivation of the model for power spectral density of spin noise from superfluid turbulence in the core of a neutron star, from first principles. The model features a spectral turnover, which depends on the dynamical response time of the superfluid and the steady-state angular velocity lag between the crust and the core of the star. In this work, we search for a spectral turnover in spin noise using the first data release of the International Pulsar Timing Array. Through Bayesian model selection, we find no evidence of a spectral turnover. Our analysis also shows that data from pulsars J1939+2134, J1024$-$0719 and J1713+0747 prefers the power-law model to the superfluid turbulence model.