In one approach, I will use high signal-to-noise radio profiles to find the spatial distribution of radio-loud plasma in the polar cap. This can be done by means of a numerical code which starts with a model of the emission geometry and predicts the observed profile; comparison with the data allows the chosen model to be accepted or rejected. The geometry of many pulsars can be explained by the simple case of dipolar magnetic fields. Using multi-frequency data on a large number of objects (taken at Arecibo by T. H. Hankins & J. M. Rankin), my colleagues and I are using this model to determine the emission geometry.
Some other pulsars, on the other hand, cannot easily be explained by dipolar geometry (such as the Crab pulsar or millisecond pulsars). To address this, I am extending the numerical code to include both relativistic effects and quadropole contributions to the magnetic field.
In another approach, I use the fact that the radio signal, if emitted low in the polar cap, will propagate through the pulsar atmosphere before escaping and being observed at earth. This propagation adds a distinctive signal to the radio pulse: it is dispersed in frequency and broadened in time. Both of these effects can be distinguished from propagation through the galactic interstellar medium, if the observations are made at high enough frequencies. In the near future I intend to use such data to determine densities and speeds within the polar cap of the Crab pulsar.
