Paper detail

Pulsar State Switching from Markov Transitions and Stochastic Resonance

Markov processes are shown to be consistent with metastable states seen in pulsar phenomena, including intensity nulling, pulse-shape mode changes, subpulse drift rates, spindown rates, and X-ray emission, based on the typically broad and monotonic distributions of state lifetimes. Markovianity implies a nonlinear magnetospheric system in which state changes occur stochastically, corresponding to transitions between local minima in an effective potential. State durations (though not transition times) are thus largely decoupled from the characteristic time scales of various magnetospheric processes. Dyadic states are common but some objects show at least four states with some transitions forbidden. Another case is the long-term intermittent pulsar B1931+24 that has binary radio-emission and torque states with wide, but non-monotonic duration distributions. It also shows a quasi-period of $38\pm5$ days in a 13-yr time sequence, suggesting stochastic resonance in a Markov system with a forcing function that could be strictly periodic or quasi-periodic. Nonlinear phenomena are associated with time-dependent activity in the acceleration region near each magnetic polar cap. The polar-cap diode is altered by feedback from the outer magnetosphere and by return currents from an equatorial disk that may also cause the neutron star to episodically charge and discharge. Orbital perturbations in the disk provide a natural periodicity for the forcing function in the stochastic resonance interpretation of B1931+24. Disk dynamics may introduce additional time scales in observed phenomena. Future work can test the Markov interpretation, identify which pulsar types have a propensity for state changes, and clarify the role of selection effects.